Mass Spectrometry Imaging: Principles and Protocols [1 ed.] 1607617455, 9781607617457, 9781607617464

Citation preview

ME T H O D S

IN

MO L E C U L A R BI O L O G Y

Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK

For other titles published in this series, go to www.springer.com/series/7651

TM

Mass Spectrometry Imaging Principles and Protocols

Edited by

Stanislav S. Rubakhin Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, IL, USA

Jonathan V. Sweedler Department of Chemistry, University of Illinois at Urbana-Champaign, Urbana, IL, USA

Editors Stanislav S. Rubakhin Beckman Institute University of Illinois at Urbana-Champaign N. Mathews Avenue 405 61801 Urbana Illinois USA [email protected]

Jonathan V. Sweedler Department of Chemistry University of Illinois at Urbana-Champaign South Mathews Avenue 63-5 600 61801 Urbana Illinois USA [email protected]

ISSN 1064-3745 e-ISSN 1940-6029 ISBN 978-1-60761-745-7 e-ISBN 978-1-60761-746-4 DOI 10.1007/978-1-60761-746-4 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2010930696 © Springer Science+Business Media, LLC 2010 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein. Printed on acid-free paper Humana Press is part of Springer Science+Business Media (www.springer.com)

Preface Why should you use the protocols described in this book? In order to characterize many biological systems, having knowledge of their chemical constituents, locations, and dynamics is important. Mass spectrometry (MS) provides unmatched capabilities for detection, characterization, and identification of analytes ranging from individual elements to complex multimolecular structures. A powerful enhancement to MS detection is the addition of spatial information; mass spectrometry imaging (MSI) combines the capabilities of modern MS with imaging. The distribution of hundreds of different analytes in a tissue can be determined in a single experiment. Unlike other imaging approaches, analyte preselection is not needed. Metabolites, peptides, proteins, and polynucleotides can be characterized. Distinct protocols have been developed for analysis of specimens in vivo, in vitro, and in situ. This book is divided into three parts. The first section contains introductory chapters on MS and MSI. Chapter 1 provides an overview of MSI and focuses on current and future trends in the field. The success of a particular MSI experiment depends on the specific MS approach used. Therefore, the second chapter describes the basic principles of mass spectrometry relevant to MSI and includes cross-references to other chapters of this volume for easier navigation. The third chapter reviews the application of MSI to the study of elemental distributions. Following these introductory chapters, there are multiple protocols that describe qualitative and quantitative measurements of endogenous metabolites and xenobiotics as well as their identification and localization. The last section includes protocols for a variety of MSI approaches developed to study peptide and protein distributions. The experimental protocols presented herein encompass most MSI approaches and technologies for samples from a wide range of biological models including plants, invertebrates, and vertebrates. The contributors to this volume include practitioners in academia, industry, and the clinic. MSI has great unmet potential to further the investigations in many disciplines, including molecular biology and neuroscience. This book is written for scientists who want to apply MSI methods to their research and would benefit from the detailed, stepby-step experimental protocols provided. Both novice and established MSI practitioners should find the volume a source of valuable methodological information. Included with the protocols are additional troubleshooting notes that highlight important nuances and potential pitfalls of the procedures outlined within each chapter. We hope you enjoy reading this volume as much as we enjoyed putting it together. Stanislav S. Rubakhin and Jonathan V. Sweedler

v

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

PART I

INTRODUCTORY REVIEWS AND TUTORIALS

1.

Imaging Mass Spectrometry: Viewing the Future . . . . . . . . . . . . . . . . . Sarah A. Schwartz and Richard M. Caprioli

3

2.

A Mass Spectrometry Primer for Mass Spectrometry Imaging . . . . . . . . . . . Stanislav S. Rubakhin and Jonathan V. Sweedler

21

3.

Imaging of Metals, Metalloids, and Non-metals by Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS) in Biological Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J. Sabine Becker and J. Susanne Becker

PART II

51

PROTOCOLS FOR MS IMAGING OF DISTRIBUTION OF S MALL M OLECULES I NCLUDING M ETABOLITES AND P HARMACEUTICALS

4.

Lipid Detection, Identification, and Imaging Single Cells with SIMS . . . . . . . Michael L. Heien, Paul D. Piehowski, Nicholas Winograd, and Andrew G. Ewing

5.

The Application and Potential of Ion Mobility Mass Spectrometry in Imaging MS with a Focus on Lipids . . . . . . . . . . . . . . . . . . . . . . . Amina S. Woods and Shelley N. Jackson

85

99

6.

Quantitative Imaging of Chemical Composition in Single Cells by Secondary Ion Mass Spectrometry: Cisplatin Affects Calcium Stores in Renal Epithelial Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Subhash Chandra

7.

Imaging MALDI Mass Spectrometry of Sphingolipids Using an Oscillating Capillary Nebulizer Matrix Application System . . . . . . . . . . . . . 131 Yanfeng Chen, Ying Liu, Jeremy Allegood, Elaine Wang, Begoña Cachón-González, Timothy M. Cox, Alfred H. Merrill, Jr., and M. Cameron Sullards

8.

Mapping Pharmaceuticals in Rat Brain Sections Using MALDI Imaging Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Yunsheng Hsieh, Fangbiao Li, and Walter A. Korfmacher

vii

viii 9.

Contents Laser Ablation Electrospray Ionization for Atmospheric Pressure Molecular Imaging Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . 159 Peter Nemes and Akos Vertes

10. Matrix-Assisted Laser Desorption/Ionization and Nanoparticle-Based Imaging Mass Spectrometry for Small Metabolites: A Practical Protocol . . . . . . 173 Yuki Sugiura and Mitsutoshi Setou 11. Cellular Imaging Using Matrix-Enhanced and Metal-Assisted SIMS . . . . . . . . 197 A.F. Maarten Altelaar and Sander R. Piersma 12. Tandem Mass Spectrometric Methods for Phospholipid Analysis from Brain Tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 Timothy J. Garrett and Richard A. Yost 13. Chemical Imaging with Desorption Electrospray Ionization Mass Spectrometry . . 231 Vilmos Kertesz and Gary J. Van Berkel 14. Mass Spectrometry Imaging of Small Molecules Using Matrix-Enhanced Surface-Assisted Laser Desorption/Ionization Mass Spectrometry (ME-SALDI-MS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Qiang Liu, Yongsheng Xiao, and Lin He 15. Preparation of Single Cells for Imaging Mass Spectrometry . . . . . . . . . . . . 253 Elena S.F. Berman, Susan L. Fortson, and Kristen S. Kulp 16. Applying Imaging ToF-SIMS and PCA in Differentiation of Tissue Types . . . . . 267 Ligang Wu, James S. Felton, and Kuang Jen J. Wu PART III

PROTOCOLS FOR MS IMAGING OF DISTRIBUTION OF PEPTIDES AND P ROTEINS

17. Direct Molecular Analysis of Whole-Body Animal Tissue Sections by MALDI Imaging Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . 285 Michelle L. Reyzer, Pierre Chaurand, Peggi M. Angel, and Richard M. Caprioli 18. MALDI Direct Analysis and Imaging of Frozen Versus FFPE Tissues: What Strategy for Which Sample? . . . . . . . . . . . . . . . . . . . . . . . . . 303 Maxence Wisztorski, Julien Franck, Michel Salzet, and Isabelle Fournier 19. On Tissue Protein Identification Improvement by N-Terminal Peptide Derivatization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 Julien Franck, Mohamed El Ayed, Maxence Wisztorski, Michel Salzet, and Isabelle Fournier 20. Specific MALDI-MSI: TAG-MASS . . . . . . . . . . . . . . . . . . . . . . . . . 339 Jonathan Stauber, Mohamed El Ayed, Maxence Wisztorski, Michel Salzet, and Isabelle Fournier

Contents

ix

21. Structurally Selective Imaging Mass Spectrometry by Imaging Ion Mobility-Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363 John A. McLean, Larissa S. Fenn, and Jeffrey R. Enders 22. Tutorial: Multivariate Statistical Treatment of Imaging Data for Clinical Biomarker Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385 Sören-Oliver Deininger, Michael Becker, and Detlev Suckau 23. Applications of MALDI-MSI to Pharmaceutical Research . . . . . . . . . . . . . 405 Brendan Prideaux, Dieter Staab, and Markus Stoeckli 24. Tissue Preparation for the In Situ MALDI MS Imaging of Proteins, Lipids, and Small Molecules at Cellular Resolution . . . . . . . . . . . . . . . . . 415 Nathalie Y.R. Agar, Jane-Marie Kowalski, Paul J. Kowalski, John H. Wong, and Jeffrey N. Agar 25. Imaging of Similar Mass Neuropeptides in Neuronal Tissue by Enhanced Resolution MALDI MS with an Ion Trap – OrbitrapTM Hybrid Instrument . . . . 433 Peter D.E.M. Verhaert, Martijn W.H. Pinkse, Kerstin Strupat, and Maria C. Prieto Conaway 26. Mass Spectrometric Imaging of Neuropeptides in Decapod Crustacean Neuronal Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451 Ruibing Chen, Stephanie S. Cape, Robert M. Sturm, and Lingjun Li 27. Mass Spectrometry Imaging Using the Stretched Sample Approach . . . . . . . . 465 Tyler A. Zimmerman, Stanislav S. Rubakhin, and Jonathan V. Sweedler Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481

Contributors JEFFREY N. AGAR • Chemistry Department and Volen Center, Brandeis University, Waltham, MA, USA NATHALIE Y.R. AGAR • Department of Neurosurgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA JEREMY ALLEGOOD • School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, GA, USA PEGGI M. ANGEL • Mass Spectrometry Research Center, Vanderbilt University, Nashville, TN, USA MICHAEL BECKER • Applications TOF-MS, Bruker Daltonik GmbH, Bremen, Germany ELENA S.F. BERMAN • Los Gatos Research, Mountain View, CA, USA BEGOÑA CACHÓN-GONZÁLEZ • Department of Medicine, University of Cambridge, Cambridge, UK M. CAMERON SULLARDS • School of Chemistry and Biochemistry, School of Biology, Georgia Institute of Technology, Atlanta, GA, USA STEPHANIE S. CAPE • School of Pharmacy and Department of Chemistry, University of Wisconsin-Madison, Madison, WI, USA RICHARD M. CAPRIOLI • Departments of Chemistry and Biochemistry, Mass Spectrometry Research Center, Vanderbilt University, Nashville, TN, USA SUBHASH CHANDRA • Cornell SIMS Laboratory, Department of Biomedical Engineering, Cornell University, Ithaca, NY, USA PIERRE CHAURAND • Mass Spectrometry Research Center, Vanderbilt University, Nashville, TN, USA RUIBING CHEN • School of Pharmacy and Department of Chemistry, University of Wisconsin-Madison, Madison, WI, USA YANFENG CHEN • School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, GA, USA TIMOTHY M. COX • Department of Medicine, University of Cambridge, Cambridge, UK SÖREN-OLIVER DEININGER • Applications TOF-MS, Bruker Daltonik GmbH, Bremen, Germany MOHAMED EL AYED • Laboratoire de Neuroimmunologie et Neurochimie Evolutives, FRE CNRS 3249, MALDI Imaging Team, Université Lille Nord de France, Université Lille 1, Villeneuve d’Ascq, France JEFFREY R. ENDERS • Department of Chemistry, Vanderbilt Institute of Chemical Biology, and Vanderbilt Institute for Integrative Biosystems Research and Education, Vanderbilt University, Nashville, TN, USA ANDREW G. EWING • Department of Chemistry, The Pennsylvania State University, University Park, PA, USA; Department of Chemistry, Göteborg University, Göteborg, Sweden JAMES S. FELTON • Chemistry, Materials, Earth, and Life Sciences (CMELS) Directorate, Lawrence Livermore National Laboratory, Livermore, CA, USA

xi

xii

Contributors

LARISSA S. FENN • Department of Chemistry, Vanderbilt Institute of Chemical Biology, and Vanderbilt Institute for Integrative Biosystems Research and Education, Vanderbilt University, Nashville, TN, USA SUSAN L. FORTSON • Los Gatos Research, Mountain View, CA, USA ISABELLE FOURNIER • Laboratoire de Neuroimmunologie et Neurochimie Evolutives, FRE CNRS 3249, MALDI Imaging Team, Université Lille Nord de France, Université Lille 1, Villeneuve d’Ascq, France JULIEN FRANCK • Laboratoire de Neuroimmunologie et Neurochimie Evolutives, FRE CNRS 3249, MALDI Imaging Team, Université Lille Nord de France, Université Lille 1, Villeneuve d’Ascq, France TIMOTHY J. GARRETT • GCRC Core Laboratory, Department of Medicine, University of Florida, Gainesville, FL, USA LIN HE • Department of Chemistry, North Carolina State University, Raleigh, NC, USA MICHAEL L. HEIEN • Assistant Professor, Department of Chemistry and Biochemistry, The University of Arizona, Tucson, AZ, USA YUNSHENG HSIEH • Merck Research Laboratories, Department of Drug Metabolism and Pharmacokinetics, Kenilworth, NJ, USA SHELLEY N. JACKSON • Structural Biology Unit, Cellular Neurobiology Branch, National Institute on Drug Abuse Intramural Research Program, National Institutes of Health, Baltimore, MD, USA VILMOS KERTESZ • Organic and Biological Mass Spectrometry Group, Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA WALTER A. KORFMACHER • Department of Drug Metabolism and Pharmacokinetics, Merck Research Laboratories, Kenilworth, NJ, USA JANE-MARIE KOWALSKI • Bruker Daltonics, Inc., Billerica, MA, USA PAUL J. KOWALSKI • Bruker Daltonics, Inc., Billerica, MA, USA KRISTEN S. KULP • Biosciences and Biotechnology Division, Lawrence Livermore National Laboratory, Livermore, CA, USA FANGBIAO LI • Department of Drug Metabolism and Pharmacokinetics, Merck Research Laboratories, Kenilworth, NJ, USA LINGJUN LI • School of Pharmacy and Department of Chemistry, University of WisconsinMadison, Madison, WI, USA QIANG LIU • Department of Pathology & Lab Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA YING LIU • School of Biology, Georgia Institute of Technology, Atlanta, GA, USA A.F. MAARTEN ALTELAAR • Biomolecular Mass Spectrometry and Proteomics Group, Utrecht University, Utrecht, The Netherlands JOHN A. MCLEAN • Department of Chemistry, Vanderbilt Institute of Chemical Biology, and Vanderbilt Institute for Integrative Biosystems Research and Education, Vanderbilt University, Nashville, TN, USA ALFRED H. MERRILL, JR. • School of Biology, School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, GA, USA PETER NEMES • W.M. Keck Institute for Proteomics Technology and Applications, Department of Chemistry, George Washington University, Washington, DC, USA PAUL D. PIEHOWSKI • Department of Chemistry, The Pennsylvania State University, University Park, PA, USA

Contributors

xiii

SANDER R. PIERSMA • OncoProteomics Laboratory, Department of Medical Oncology, VUmc-Cancer Center Amsterdam, VU University Medical Center Amsterdam, Amsterdam, The Netherlands MARTIJN W.H. PINKSE • Kluyver Laboratory, Department of Biotechnology, Netherlands Proteomics Center, Delft University of Technology, Delft, The Netherlands BRENDAN PRIDEAUX • Analytical Sciences, Novartis Institutes for BioMedical Research, Basel, Switzerland MARIA C. PRIETO CONAWAY • Thermo Fisher Scientific, San Jose, CA, USA MICHELLE L. REYZER • Mass Spectrometry Research Center, Vanderbilt University, Nashville, TN, USA STANISLAV S. RUBAKHIN • Department of Chemistry and Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, IL, USA J. SABINE BECKER • Central Division of Analytical Chemistry, Forschungszentrum Jülich, Jülich, Germany MICHEL SALZET • Laboratoire de Neuroimmunologie et Neurochimie Evolutives, FRE CNRS 3249, MALDI Imaging Team, Université Lille Nord de France, Université Lille 1, Villeneuve d’Ascq, France SARAH A. SCHWARTZ • David H. Murdock Research Institute, North Carolina Research Campus, Kannapolis, NC, USA MITSUTOSHI SETOU • Mitsubishi Kagaku Institute of Life Sciences, Tokyo, Japan; Department of Molecular Anatomy, Hamamatsu University School of Medicine, Shizuoka, Japan DIETER STAAB • Novartis Institutes for BioMedical Research, Basel, Switzerland JONATHAN STAUBER • Laboratoire de Neuroimmunologie et Neurochimie Evolutives, FRE CNRS 3249, MALDI Imaging Team, Université Lille Nord de France, Université Lille 1, Villeneuve d’Ascq, France MARKUS STOECKLI • Novartis Institutes for BioMedical Research, Basel, Switzerland KERSTIN STRUPAT • Thermo Fisher Scientific GmbH, Bremen, Germany ROBERT M. STURM • School of Pharmacy and Department of Chemistry, University of Wisconsin-Madison, Madison, WI, USA DETLEV SUCKAU • Applications TOF-MS, Bruker Daltonik GmbH, Bremen, Germany YUKI SUGIURA • Department of Bioscience and Biotechnology, Tokyo Institute of Technology, Kanagawa, Japan; Mitsubishi Kagaku Institute of Life Sciences, Tokyo, Japan J. SUSANNE BECKER • Aeropharm GmbH, Rudolstadt, Germany JONATHAN V. SWEEDLER • Department of Chemistry and Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, IL, USA GARY J. VAN BERKEL • Organic and Biological Mass Spectrometry Group, Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA PETER D.E.M. VERHAERT • Kluyver Laboratory, Department of Biotechnology, Netherlands Proteomics Center, Delft University of Technology, Delft, The Netherlands; Laboratory of Molecular Cell Biology, VIB, Flemish Institute of Biotechnology, Leuven, Belgium; BioMedical Research Center, University of Hasselt, Diepenbeek, Belgium AKOS VERTES • W.M. Keck Institute for Proteomics Technology and Applications, Department of Chemistry, George Washington University, Washington, DC, USA ELAINE WANG • School of Biology, Georgia Institute of Technology, Atlanta, GA, USA NICHOLAS WINOGRAD • Department of Chemistry, The Pennsylvania State University, University Park, PA, USA

xiv

Contributors

MAXENCE WISZTORSKI • Laboratoire de Neuroimmunologie et Neurochimie Evolutives, FRE CNRS 3249, MALDI Imaging Team, Université Lille Nord de France, Université Lille 1, Villeneuve d’Ascq, France JOHN H. WONG • Chemistry Department and Volen Center, Brandeis University, Waltham, MA, USA AMINA S. WOODS • Structural Biology Unit, Cellular Neurobiology Branch, National Institute on Drug Abuse Intramural Research Program, National Institutes of Health, Baltimore, MD, USA KUANG JEN J. WU • Bioscience and Biotechnology Division, Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, CA, USA LIGANG WU • Seagate Technology US, Fremont, CA, USA YONGSHENG XIAO • Department of Chemistry, University of California, Riverside, CA, USA RICHARD A. YOST • Department of Chemistry, University of Florida, Gainesville, FL, USA TYLER A. ZIMMERMAN • Department of Chemistry and Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, IL, USA

Part I Introductory Reviews and Tutorials

Chapter 1 Imaging Mass Spectrometry: Viewing the Future Sarah A. Schwartz and Richard M. Caprioli Abstract Imaging mass spectrometry (IMS) technology is an effective tool that is able to assess complex molecular mixtures in cells, tissues, or other sample types with high chemical specificity, allowing concurrent analysis of a variety of molecular species in a wide mass range, from small metabolites to large macromolecules such as proteins. Simultaneous localization of molecules, detection of post-translational modifications, and relative quantitative information can be obtained in a single experiment. Images generated by MS are unique because they are derived from direct molecular measurements and do not rely on targetspecific reagents such as antibodies. Thus, the ability to map spatial distributions coupled with the mass accuracy and chemical specificity for MS-based detection makes IMS an effective discovery tool. Further structural assessment of compounds, including MS/MS fragmentation analysis, can be utilized in an imaging experiment to achieve accurate molecular identifications. Key words: Imaging mass spectrometry, secondary ion mass spectrometry, matrix-assisted laser desorption ionization, desorption electrospray ionization, brain.

1. Introduction The complexity of cells, tissues, organs, and whole systems is defined not only by the myriad of molecular events that occur but also by their distribution in space and time. Imaging mass spectrometry (IMS) technology provides a tool to assess these events with high chemical specificity, allowing concurrent analysis of a variety of molecular species in a wide mass range, from small metabolites to large macromolecules such as proteins. Simultaneous localization of molecules, detection of post-translational modifications, and relative quantitative information can be obtained in a single experiment. Although images S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_1, © Springer Science+Business Media, LLC 2010

3

4

Schwartz and Caprioli

generated by MS are similar in appearance to more traditional methodologies such as chemical staining, immunohistochemistry, and radiochemistry, they are unique because they are derived from direct molecular measurements and do not rely on target-specific reagents such as antibodies. Thus, the ability to map spatial distributions coupled with the mass accuracy and chemical specificity for MS-based detection makes IMS an effective discovery tool. Further structural assessment, including MS/MS fragmentation analysis, can be utilized for molecular identification (1). IMS experiments may be performed using one of several MS ionization techniques: secondary ion mass spectrometry (SIMS), laser desorption/ionization (LDI), desorption electrospray ionization (DESI), and matrix-assisted laser desorption/ionization (MALDI) (2). In general, these techniques offer complementary capabilities (3). SIMS imaging is favored for higher spatial resolution imaging over a low mass range ( m/z

d2 Reflecto r

m/z

Linear detector

DriŌ zone

Ion source

m ode

Reflector detector

Ion mirror ON

Fig. 2.8. Schematics of co-axial geometry time-of-flight (TOF) mass analyzer operation in linear and reflector modes. A simplified equation describing the TOF process is presented in the insert. m = mass of the ion, V = acceleration voltage, d = length of flight path, t = time from the moment of ion acceleration to the detection event.

A Mass Spectrometry Primer for MSI

37

ions are extracted from the ion source and unidirectionally accelerated by short pulses of electrostatic field, entering and moving in a drift space containing no field; all ions are accelerated with the same kinetic energy. Thus, lighter ions of the same charge will move faster than heavier ones and, therefore, come to the detector earlier; obviously, separation takes place in space and detection in time. One of the drawbacks of the linear mode is that all the ions initially do not have the exact same position and velocity, so that there is a spread in ion arrival times at the detector. This spatial distribution leads to formation of broad, lower amplitude signals at the detector, resulting in reduced resolution and at times lower sensitivity of detection. There are two approaches used in TOF mass analyzers to improve resolution. One is delayed extraction (delayed injection into the mass analyzer, also called pulsed ion extraction) of the ions formed in the MALDI ion source. This delay reduces the energy spread of the same m/z ions. Delayed extraction parameters can be adjusted by the operator, with a longer delay time needed for improved detection of larger molecules. Another approach utilizes reflecting ion optics (reflector or reflectron mode) such as ion mirrors – a set of evenly spaced electrodes encompassing space on the ion path (Fig. 2.8). A single linear electric field with higher potential energy than the source potential is formed around each electrode. Ions are flying in these fields and are reflected (repulsed). As a result, resolution improves due to two factors. First, there is an increase in ion path length and thus, a greater distance between packets of ions; second, there is a reduction in the spread of kinetic energies of different particles of the same m/z. More energetic ions will travel longer paths in the field space than lower kinetic energy ions; therefore, they will be focused as they leave the ion mirror area, arriving at the detector in a more temporally compact packet. Both approaches are typically used when metabolites, peptides, and small- and mediumsized proteins are analyzed. However, the linear mode of operation with delayed extraction remains the preferred option for analysis of large molecules and molecular aggregates. TOF analyzers are common in MSI applications because of their speed of operation and wide m/z range. They allow analysis of large singly charged molecular ions produced by MALDI (see Chapters 7–11, 14–20, 22–24, 26, and 27). 2.2.2.3. Sector Instruments

Magnetic sector and electrostatic sector mass analyzers are well suited for operation with continuous ion sources; the trajectories of moving ions are curved by forces developed by the electric or magnetic fields (Fig. 2.9). The extent of this curvature depends on an ion’s m/z. Sector analyzers can be used to monitor a single ion with high resolution. A narrow slit is installed between the detector and the ion analyzer; the position of the slit determines

Rubakhin and Sweedler

z m/

rating Accele ge a lt o V (V)

(r) Electric sector r = 2v/E Magnetic sector r = mv/BZ

/z >m

Ion e Sourc

Ra diu s

38

tor(S) Detec Centrifugal force = Centripetal force Kinetic energy = Potential energy Direction focusing m/Z = B2r2/2V m/Z = Er/v2

Fig. 2.9. General schematic of a sector mass analyzer. Ions extracted from the ion source are accelerated by an electrostatic field (accelerating potential, V ) and enter the sector analyzer with velocity, v. Electric (electric flux density, E ) or magnetic (magnetic flux density, B ) fields bend the trajectory of the ions into curved paths with radius, r. Trajectories of ions with larger m/z are affected more than smaller ones. An illustration of the direction-focusing ion beam approach in a magnetic sector mass analyzer is shown in the insert. Due to the dependence of the radius of an ion’s trajectory on its kinetic energy (E ) in the electrostatic sector mass analyzer and on its momentum (mv) in the magnetic sector mass analyzer, the systems are also referred to as ion energy and ion momentum filters.

which ion is detected. A narrower slit improves mass resolution but decreases sensitivity. Mass resolution is also dependent on the cross section of the incoming ion beam, the m/z ion kinetic energy spread, and the radius of ion trajectory. Different m/z ions can be recorded simultaneously by using multiple detectors (or a detector array). Due to fast ion transmission and low level of interaction between ions in the beam (e.g., minimal space-charge effects leading to ion–ion repulsions), sector analyzers are capable of quantitative measurements. Ions approach the sector analyzer as focused or defocused beams. The latter can be refocused with a direction-focusing approach using a magnetic sector mass analyzer (Fig. 2.9, inset). Electrostatic mass analyzers are efficient ion kinetic energy filters whereas magnetic sector analyzers are capable of filtering ions with differing momentum (Fig. 2.9). Therefore, hybrid instruments combining these two mass analyzers enable double focusing and may achieve 100,000 resolution. Double-focusing instruments have been employed to image the distribution of elements by LA-ICP MSI (see Chapter 3). Magnetic sector SIMS instruments (see Chapter 6), like the CAMECA IMS 7f, are capable of distinguishing such ions as 56 Fe and 28 Si2 and allow direct ion microscopy (stigmatic mode imaging) and scanning microprobe mode imaging.

A Mass Spectrometry Primer for MSI

The quadrupole analyzer is compatible with continuous ion sources such as DESI. Although the upper m/z range of the quadrupole is not high, neither is DESI well suited for desorbing high molecular weight analytes and so the two approaches work well together. The quadrupole, as its name suggests, consists of four precisely aligned metal hyperbolic or cylindrical rods (Fig. 2.10). Superimposed direct current (DC) and oscillating radiofrequency (RF) electric fields are used to create conditions where ions of only a certain m/z (typically a 1 m/z mass transmission window) will have a stable trajectory inside the device and therefore pass through it. Other ions of lower and higher m/z will leave the analyzer prematurely or collide with the rods and skeleton. Therefore, the quadrupole mass analyzer is a form of mass filter. The ion path in the quadrupole starts as circular transforms to complex spiral-like propagation inside the field. Depending on the analyzer’s design, stable trajectories for ions of a particular m/z can be achieved in an oscillating electric field by setting the appropriate RF frequency and RF and DC voltage amplitudes. Importantly, simultaneous change of DC and RF voltage ampli-

A

+ (U + Vcosωt) > m/z

m/z

– (U + Vcosωt)

> m/z

Dc potential: U RF potential: Vcosωt

g atin eler Acc age volt rce sou

Ion

– (U + Vcosωt)


10 cm2 . Ions strike the detector’s surface inside an individual channel, thereby inducing secondary electron and photon emission. The efficiency of this process depends on the kinetic energy of incoming ions, incident angle of impact, and detector surface properties. Secondary electrons continue to move toward the detector and strike tunnel walls, again inducing more electrons. This effect can multiply the number of ions by more than a million fold. The process continues along the channel until electrons reach a conductor, which transmits this electric current to amplifiers and further signal processing. Detectors operate at vacuum conditions usually below 10–5 torr. However, new technologies for generation of curved channels in channel electron multipliers have resulted in an ion detector that is operational at 10–2 torr (http://www.detechinc.com/em/quad.htm). Importantly, the ion detectors of some mass spectrometers can be replaced/updated, translating into significant improvements in detection capabilities. Why is it important to understand how the detector works? Many mass spectrometers have the option of adjusting detector sensitivity. Increasing the sensitivity may help to observe more compounds and aid in the detection of low-abundance analytes. Unfortunately, detector operation at a higher sensitivity setting can increase the noise level and reduce the lifetime of the detector. In addition, if the detector is set at a higher sensitivity and intense signals are present, this can degrade detector performance. High spatial resolution MS imaging requires significant dynamic range

A Mass Spectrometry Primer for MSI

45

and often the detector is optimized for low analyte levels at each probed spot. Thus, detector performance can degrade during the hundreds of thousands of acquisitions that take place when creating a number of larger ion images. What is next? The signals produced by detectors are digitized, processed, and converted to mass spectra. MS imaging experiments may generate thousands of mass spectra containing information on hundreds of signals acquired from specimens. Obviously, it is not possible to check individual mass spectrum quality. Therefore, data analysis in MSI experiments is mostly done using final ion images. Such analysis may generate false-positive signal detection due to a variety of factors, including issues as shown in Fig. 2.13. Different baseline correction, denoising, and intel-

Fig. 2.13. Factors complicating analysis of MSI experiment results. (a) Formation of alkali metal adducts. Sodium and potassium adducts are marked on the mass spectrum. These can occur in a region-specific manner in an MSI image. (b) Detector saturation. Mass spectra acquired with different laser fluencies (the total energy per unit area). Note lost resolution on the major peak and improved detection of lower intensity peaks as a result of fluency increase (lower trace). (c) Curved baseline shape, chemical, and/or digitization noise may produce false-positive peak detection during the automatic peak picking process. Mass spectrum shows an elevated baseline and high level of chemical noise (left inset) in the lower m/z range, mixed chemical and digitization noises in the middle range (central inset), and only digitization noise in the higher m/z region.

46

Rubakhin and Sweedler

ligent peak picking algorithms are implemented for batch mass spectra processing in MSI (40, 41) (see Chapters 1, 22, and 27). Nevertheless, a random check of several of the original mass spectra is an important step during data analysis. 2.2.4. Calibration

Mass spectrometers are complex systems and many parameters can be optimized according to experimental requirements. These adjustments, as well as possible drift in the calibration of different components, may lead to measurement of analyte m/z with a systematic error. Therefore, calibration of the mass scale is often performed using standards, often more than three; this compensates for these deleterious effects and results in a more accurate m/z scale. During automatic or manual calibration, standards with known masses and predictable charge are measured, differences between determined mass and true mass are found, and a set of calibration constants are created. These constants become part of the m/z calibration during subsequent measurements performed with the same instrumental settings with unknown analytes. Internal, external, or mass defect-based calibration can be performed (42). External calibration is done with mixtures of standards measured separately from analyzed unknown samples. In contrast, internal calibration requires the presence of calibrants in the sample. Importantly, if the m/z of endogenous analytes are known, they can be used to calibrate the mass spectrometer. This type of calibration is preferable because no increase in complexity of sample occurs and influences from different parameters, such as sample thickness and analyte environment, are accounted for. As an example, trypsin autocleavage molecular ions are used for calibration in proteomics research (43), as is mass defect-based calibration, where peptides of enzymatically cleaved proteins are used for calibration and protein identifications (42, 44). The latter approach is also useful for determination and subsequent elimination of nonpeptide species from peak lists. Calibrant selection is governed by several requirements, including the nature of targeted analytes, investigated m/z range, type of mass spectrometer, and calibration approach. Unique, exogenous calibrants are preferable for internal calibration when endogenous calibrants are not available. Using analytes that can be endogenously present in an investigated sample for calibration is discouraged due to the possibility of ambiguous experimental outcomes, as well as accidental contamination of other samples. High-quality calibration is achieved with calibrants that bracket the mass of the analyte of interest. After their acquisition, individual mass spectra can be recalibrated employing different calibration profiles. However, recalibration of complex MSn spectra, or an entire set of mass spectra obtained during an MS imaging experiment, can be difficult. Therefore, it is important to verify calibration before performing an imaging experiment using the appropriate standards.

A Mass Spectrometry Primer for MSI

47

3. Conclusions Mass spectrometry provides a set of tools with versatile and powerful figures of merit. When used under the control of appropriate operating protocols, many MS-platforms can be used for MSI. The large number of data-points in an image, and the complex matrix of working directly with tissue samples, can place severe performance constraints on an MSI experiment. Nonetheless, the rapid increase in the performance of ion sources, mass analyzers, and detectors has led to a new generation of sensitive, highresolution, and relatively easy-to-use mass spectrometers. These are now providing unmatched capabilities for chemical imaging. The mature mass spectrometry field provides biologists with a variety of novel imaging modalities that are appropriate for a number of applications. While there is no single MS-based tool that is ideal for all analytes, the assortment of instruments now available allows a broad range of samples to be probed. In the following chapters, specific protocols are described in much greater detail.

Acknowledgments The authors would like to thank Kevin Tucker for helpful discussion of the chapter and Stephanie Baker for her help with the manuscript preparation. This material is based on work supported by Award No. P30DA018310 from the National Institute on Drug Abuse (NIDA). The content is solely the responsibility of the authors and does not necessarily represent the official views of NIDA or the National Institutes of Health. References 1. Burlingame, A. L. (2005) Mass spectrometry: modified proteins and glycoconjugates, Methods in Enzymolology. Elsevier Academic Press, Amsterdam, Boston, MA. 2. Chance, M. (ed.) (2008) Mass Spectrometry Analysis for Protein–Protein Interactions and Dynamics. John Wiley & Sons, Hoboken, NJ. 3. Downard, K. (ed.) (2007) Mass Spectrometry of Protein Interactions. Wiley Interscience, Hoboken, NJ. 4. Lipton, M. S., Páya-Tolic, L. (2009) Mass Spectrometry of Proteins and Pep-

5. 6. 7. 8.

tides: Methods and Protocols. Humana Press, Springer, distributor, New York, NY, London. Matthiesen, R. (ed.) (2007) Mass Spectrometry Data Analysis in Proteomics. Humana Press, Totowa, NJ. Murphy, R. C. (1993) Mass Spectrometry of Lipids. Plenum Press, New York, NY. Sparkman, O. D. (2006) Mass Spec Desk Reference. 2nd ed. Global View Publishing, Pittsburgh, PA. Wanner, K. T., Höfner, G. (eds.) (2007) Mass Spectrometry in Medicinal Chemistry:

48

9. 10.

11. 12.

13.

14.

15.

16.

17.

18.

19.

Rubakhin and Sweedler Applications in Drug Discovery. WileyVCH, John Wiley (distributor), Weinheim, Chichester. McLafferty, F. W., Turecek, F. (1993) Interpretation of Mass Spectra. University Science Books, Mill Valley, CA. Becker, J. S. (2007) Inorganic Mass Spectrometry: Principles and Applications. John Wiley & Sons, Chichester, Englandm Hoboken, NJ. De Podesta, M. (2001) Understanding the Properties of Matter. Taylor & Francis, London, New York, NY. Holscher, D., Shroff, R., Knop, K., Gottschaldt, M., Crecelius, A., Schneider, B., Heckel, D. G., Schubert, U. S., Svatos, A. (2009) Matrix-free UV-laser desorption/ionization (LDI) mass spectrometric imaging on the single-cell level: distribution of secondary metabolites of Arabidopsis thaliana and Hypericum species. Plant J, 60, 907–918. Liu, Q., Guo, Z., He, L. (2007) Mass spectrometry imaging of small molecules using desorption/ionization on silicon. Anal Chem, 79, 3535–3541. Van Berkel, G. J., Kertesz, V., Koeplinger, K. A., Vavrek, M., Kong, A. N. (2008) Liquid microjunction surface sampling probe electrospray mass spectrometry for detection of drugs and metabolites in thin tissue sections. J Mass Spectrom, 43, 500–508. Northen, T. R., Yanes, O., Northen, M. T., Marrinucci, D., Uritboonthai, W., Apon, J., Golledge, S. L., Nordstrom, A., Siuzdak, G. (2007) Clathrate nanostructures for mass spectrometry. Nature, 449, 1033–1036. Yanes, O., Woo, H. K., Northen, T. R., Oppenheimer, S. R., Shriver, L., Apon, J., Estrada, M. N., Potchoiba, M. J., Steenwyk, R., Manchester, M., Siuzdak, G. (2009) Nanostructure initiator mass spectrometry: tissue imaging and direct biofluid analysis. Anal Chem, 81, 2969–2975. Willingham, D., Kucher, A., Winograd, N. (2008) Molecular depth profiling and imaging using cluster ion beams with femtosecond laser postionization. Appl Surf Sci, 255, 831–833. Sakamoto, T., Koizumi, M., Kawasaki, J., Yamaguchi, J. (2008) Development of a high lateral resolution TOF-SIMS apparatus for single particle analysis. Appl Surf Sci, 255, 1617–1620. Wucher, A., Cheng, J., Zheng, L., Winograd, N. (2009) Three-dimensional depth profiling

20.

21.

22. 23.

24.

25.

26.

27.

28.

29.

30.

31.

of molecular structures. Anal Bioanal Chem, 393, 1835–1842. Tanaka, K. (2003) The origin of macromolecule ionization by laser irradiation (Nobel lecture). Angew Chem Int Ed Engl, 42, 3860–3870. Hillenkamp, F., Karas, M. (1990) Mass spectrometry of peptides and proteins by matrix-assisted ultraviolet laser desorption/ionization. Methods Enzymol, 193, 280–295. Fenn, J. (2002) Electrospray ionization mass spectrometry: how it all began. J Biomol Tech, 13, 101–118. Cornett, D. S., Reyzer, M. L., Chaurand, P., Caprioli, R. M. (2007) MALDI imaging mass spectrometry: molecular snapshots of biochemical systems. Nat Methods, 4, 828–833. Kaletas, B. K., van der Wiel, I. M., Stauber, J., Guzel, C., Kros, J. M., Luider, T. M., Heeren, R. M. (2009) Sample preparation issues for tissue imaging by imaging MS. Proteomics, 9, 2622–2633. Seetharaman, V., Royston, G. (2007) Quantitative detection of metabolites using matrixassisted laser desorption/ionization mass spectrometry with 9-aminoacridine as the matrix. Rapid Commun Mass Spectrom, 21, 2072–2078. Rachal, L. V.-S., David, M. H. (2002) 9-Aminoacridine as a matrix for negative mode matrix-assisted laser desorption/ionization. Rapid Commun Mass Spectrom, 16, 1575–1581. Edwards, J. L., Kennedy, R. T. (2005) Metabolomic analysis of eukaryotic tissue and prokaryotes using negative mode MALDI time-of-flight mass spectrometry. Anal Chem, 77, 2201–2209. Liu, Y., Sun, X., Guo, B. (2003) Matrixassisted laser desorption/ionization timeof-flight analysis of low-concentration oligonucleotides and mini-sequencing products. Rapid Commun Mass Spectrom, 17, 2354–2360. Romanova, E. V., Rubakhin, S. S., Sweedler, J. V. (2008) One-step sampling, extraction, and storage protocol for peptidomics using dihydroxybenzoic acid. Anal Chem, 80, 3379–3386. Takats, Z., Wiseman, J. M., Gologan, B., Cooks, R. G. (2004) Mass spectrometry sampling under ambient conditions with desorption electrospray ionization. Science, 306, 471–473. Manicke, N. E., Wiseman, J. M., Ifa, D. R., Cooks, R. G. (2008) Desorption electrospray ionization (DESI) mass

A Mass Spectrometry Primer for MSI

32.

33.

34.

35.

36.

37.

spectrometry and tandem mass spectrometry (MS/MS) of phospholipids and sphingolipids: ionization, adduct formation, and fragmentation. J Am Soc Mass Spectrom, 19, 531–543. Wiseman, J. M., Ifa, D. R., Venter, A., Cooks, R. G. (2008) Ambient molecular imaging by desorption electrospray ionization mass spectrometry. Nat Protoc, 3, 517–524. Lundstrom, S. L., D′ Alexandri, F. L., Nithipatikom, K., Haeggstrom, J. Z., Wheelock, A. M., Wheelock, C. E. (2009) HPLC/MS/MS-based approaches for detection and quantification of eicosanoids. Methods Mol Biol, 579, 161–187. Houjou, T., Yamatani, K., Nakanishi, H., Imagawa, M., Shimizu, T., Taguchi, R. (2004) Rapid and selective identification of molecular species in phosphatidylcholine and sphingomyelin by conditional neutral loss scanning and MS3 . Rapid Commun Mass Spectrom, 18, 3123–3130. Iavarone, A. T., Jurchen, J. C., Williams, E. R. (2001) Supercharged protein and peptide ions formed by electrospray ionization. Anal Chem, 73, 1455–1460. Shi, S. D., Hendrickson, C. L., Marshall, A. G. (1998) Counting individual sulfur atoms in a protein by ultrahighresolution Fourier transform ion cyclotron resonance mass spectrometry: experimental resolution of isotopic fine structure in proteins. Proc Natl Acad Sci U S A, 95, 11532–11537. Shi, S. D. H., Drader, J. J., Hendrickson, C. L., Marshall, A. G. (1999) Fourier transform ion cyclotron resonance mass spectrometry in a high homogeneity 25 tesla resis-

38.

39.

40.

41.

42.

43.

44.

49

tive magnet. J Am Soc Mass Spectrom, 10, 265–268. Schaub, T. M., Hendrickson, C. L., Horning, S., Quinn, J. P., Senko, M. W., Marshall, A. G. (2008) High-performance mass spectrometry: Fourier transform ion cyclotron resonance at 14.5 tesla. Anal Chem, 80, 3985–3990. Makarov, A., Denisov, E., Kholomeev, A., Balschun, W., Lange, O., Strupat, K., Horning, S. (2006) Performance evaluation of a hybrid linear ion trap/orbitrap mass spectrometer. Anal Chem, 78, 2113–2120. Jardin-Mathe, O., Bonnel, D., Franck, J., Wisztorski, M., Macagno, E., Fournier, I., Salzet, M. (2008) MITICS (MALDI Imaging Team Imaging Computing System): a new open source mass spectrometry imaging software. J Proteomics, 71, 332–345. Broersen, A., van Liere, R., Altelaar, A. F., Heeren, R. M., McDonnell, L. A. (2008) Automated, feature-based image alignment for high-resolution imaging mass spectrometry of large biological samples. J Am Soc Mass Spectrom, 19, 823–832. Hjerno, K., Hojrup, P. (2007) Calibration of matrix-assisted laser desorption/ionization time-of-flight peptide mass fingerprinting spectra. Methods Mol Biol, 367, 49–60. Luo, Q., Nieves, E., Kzhyshkowska, J., Angeletti, R. H. (2006) Endogenous transforming growth factor-beta receptormediated Smad signaling complexes analyzed by mass spectrometry. Mol Cell Proteomics, 5, 1245–1260. Wolski, W. E., Farrow, M., Emde, A. K., Lehrach, H., Lalowski, M., Reinert, K. (2006) Analytical model of peptide mass cluster centres with applications. Proteome Sci, 4, 18.

Chapter 3 Imaging of Metals, Metalloids, and Non-metals by Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS) in Biological Tissues J. Sabine Becker and J. Susanne Becker Abstract The determination of the localization and distribution of essential and beneficial metals (e.g., Cu, Fe, Zn, Mn, Co, Ti, Al, Ca, K, Na, Cr and others), toxic metals (like Cd, Pb, Hg, U), metalloids (e.g., As, Se, Sb), and non-metals (such as C, S, P, Cl, I) in biological tissues is a challenging task for life science studies. Over the past few years, the development and application of mass spectrometric imaging (MSI) techniques for elements has been rapidly growing in the life sciences in order to investigate the uptake and the transport of both essential and toxic metals in plant and animal sections. Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) is a very sensitive and efficient trace, surface, and isotopic analytical technique for biological samples. LA-ICP-MS is increasingly utilized as an elemental mass spectrometric technique using double-focusing sector field (LA-ICP-SFMS) or quadrupole mass spectrometers (LA-ICP-QMS) to produce images of detailed regionally specific element distributions in thin biological tissue sections. Nowadays, MSI studies focus on brain research for studying neurodegenerative diseases such as Alzheimer’s or Parkinson’s, stroke, or tumor growth, or for the imaging of cancer biomarkers in tissue sections. The combination of the mass spectrometry imaging of metals by LA-ICP-MS with proteomics using biomolecular mass spectrometry (such as MALDI-MS or ESI-MS) to identify metal-containing proteins has become an important strategy in the life sciences. Besides the quantitative imaging of metals, nonmetals and metalloids in biological tissues, LA-ICP-MS has been utilized for imaging metal-containing proteins in a 2D gel after electrophoretic separation of proteins. Recent progress in applying LA-ICPMS in life science studies will be reviewed including the imaging of thin slices of biological tissue and applications in proteome analysis in combination with MALDI/ESI-MS to analyze metal-containing proteins. Key words: Biological tissues, mass spectrometric imaging, laser ablation inductively coupled plasma mass spectrometry, metal distribution, non-metal images, quantification, Se, gel electrophoresis.

S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_3, © Springer Science+Business Media, LLC 2010

51

52

Sabine Becker and Susanne Becker

1. Introduction Inorganic mass spectrometry has had a significant impact on chemical analysis, especially since the adoption of inductively coupled plasma mass spectrometry (ICP-MS) for the determination of element concentrations down to the trace and ultratrace level and for isotope ratio measurements on biological samples due to its very high sensitivity and low detection limits (1–3). The direct analysis of biological samples without timeconsuming sample preparation became possible by coupling a laser ablation (LA) device to ICP-MS, thus enabling information on spatial distributions of elements to be obtained (4, 5). LA-ICP-MS is today one of the most important surface mass spectrometric techniques in the life sciences for quantitative studies of metal distribution in biological tissues (human brain, animal, or plant). Compared to alternative methods in biological and clinical research for the visualization of metal distribution in tissues such as specific chemical staining, immunohistochemical staining (tags) techniques, or radiolabels for visualizing and identifying metal and molecular tags, LA-ICP-MS provides multielement capability; high sensitivity and low detection limits for most metals, metalloids and several non-metals. X-ray spectroscopic techniques for biological tissues (6), scanning electron microscopy with energy-dispersive X-ray analysis (SEM-EDX), proton-induced X-ray emission (PIXE) and autoradiography are often not sensitive enough for trace metal imaging. The application of an X-ray fluorescence nanoprobe using the Synchrotron Radiation Facility allows highly spatially resolved metal images of tissues, but the equipment and measurements are very expensive (7). Other surface inorganic mass spectrometric techniques, such as secondary ion mass spectrometry (SIMS), have been established not only for the imaging of metals but also for studying the distribution of small molecules in biological systems (8–16). Different applications of SIMS for mass spectrometric imaging are described in detail in Chapters 4, 6, 11, 15 and 16. In spite of its advantages, issues occur in SIMS if suitable matrix-matched reference materials are not available. Consequently, quantification of the analytical data is difficult due to inherent matrix effects that can be many orders of magnitude. Compared to SIMS, matrix effects are significantly lower (factor: 2–3) in LA-ICP-MS and the detection limits of imaging LA-ICP-MS on biological tissues observed at the sub-microgram per gram level are in general better (1). In addition, LA-ICP-MS possesses superior features for the quantification of analytical data in quite different systems and can be employed as a powerful and sensitive imaging technique

Imaging of Metals, Metalloids, and Non-metals

53

to produce images of detailed regionally specific element distributions in thin tissue sections of different sizes (17–21). As the most frequently used mass spectrometric imaging technique for biological tissues, matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) (22–25) is well established for the imaging of small molecules and large biomolecules up to the m/z range over 100,000 Da within biological systems (8, 22). The application of MALDI-IMS as an MS-based imaging technique has grown rapidly, enhanced by the production of commercial instrumentation and devices for sample preparation, data acquisition, and analysis (24) described in several chapters of this book. The determination of element distributions (imaging or mapping) from biological sample surfaces, such as on thin sections of biological tissues and also from nanobioelectronic devices, is of increasing relevance for a multitude of life science studies and also the development of biosensors. The analysis of essential metals (such as Cu, Zn, Fe, Ni, Mn, Mo, Mg, Ca, Na, K, Cr, B and others), metalloids (e.g., Se, As, Sb), or non-metals (like C, S, O, N, P, I, Cl and S), which are of vital importance in biological systems, is an issue in modern life science studies. Whereas a deficit of essential elements can result in various deficiency diseases, too high concentrations of these elements can be toxic. On the other hand, heavy toxic metals (like Cd, As, Cr (VI), Pb, Tl, Hg, U, Th and others) are important from the public health viewpoint and, in the low concentration range, can influence biological processes in living organisms and thus cause various illnesses. For example, the heavy metal lead (Pb2+ ) may be substituted for Ca2+ in Cadependent biological processes involved in synaptic transmission that are essential for learning, memory, growth and differentiation of nerve cells and motor function (26). Since all nutrient elements and toxic metals are in general inhomogeneously distributed in biological tissues, imaging studies require powerful analytical techniques with both good spatial resolution and high signal/noise ratio (low detection limits). As a mass spectrometric imaging technique, LA-ICP-MS enables the distribution of metals, metalloids and non-metals and isotopes to be measured (5, 17, 27). In the past, LA-ICP-MS was developed as the method of choice for the imaging of essential metals (such as Cu, Zn, Pb, Th and U – often at trace concentration levels), metalloids (Se) and selected non-metals (P, S, C, I) from microtome cut thin tissue sections (optimum at 20 µm thickness) of biological tissues (e.g., rat or human brains) (5, 17, 27–30). Advantages of LA-ICP-MS include high sample throughput, high sensitivity and accuracy and precision of the analytical data. In addition, no charging-up effects on surfaces occurs and fewer matrix effects allow straightforward quantification of analytical data. In general quantification can be performed

54

Sabine Becker and Susanne Becker

using standard reference samples of the same matrix composition. If no suitable certified standard reference materials for quantification procedures are available, alternative calibration strategies have been developed in the author’s laboratory. If no reference material is available for quantitative LA-ICP-MS analysis on biological tissues, matrix-matched laboratory standards can be prepared and employed (19, 31, 32). However, solution-based calibration also has been created as an alternative quantification procedure by inserting a micronebulizer into the laser ablation chamber (31). By means of mass spectrometric imaging analysis using LA-ICP-MS, inhomogeneous (often layered), site-specific metal distributions can be obtained in tissue sections (as demonstrated, for example, in human brain tissues for the hippocampus) or in tumor-infected regions or control brain. Furthermore, LAICP-MS has been utilized for a fast screening of 1D and 2D gels for the detection of metals, metalloids and non-metals in protein bands or protein spots separated by 1D and 2D gel electrophoresis (33–36). This chapter reviews the progress and applications of LA-ICPMS with and without a collision cell in comparison to the more sensitive double-focusing sector field LA-ICP-MS (21, 37) when used for imaging analysis from biological tissues for life science studies.

2. LA-ICP-MS Instrumentation For mass spectrometric imaging studies by LA-ICP-MS, various commercial laser ablation systems are available on the analytical market mostly using a Nd:YAG laser (e.g., from NewWave, Fremont, CA, USA, or from CETAC Technologies, Omaha, NE, USA working at wavelengths of 266 and 213 nm). The development of different laser ablation cells for LA-ICP-MS from quite different materials including the imaging of biological tissues and also gels is described in the literature (1). In commercial laser ablation cells, tissues with a maximum sample size of about 20×20 mm can be analyzed. Larger laser ablation chambers (so-called SuperCellTM ) have been designed by NewWave Research and also CETAC to analyze large thin sections of tissues as illustrated on a human brain hemisphere in Section 4 of this chapter, as well as 2D gels from the gel electrophoretic separation of proteins (and also for thin petrographic sections of geological samples and layered materials). For LA-ICP-MS measurements of dried biological tissue the laser ablation is normally performed at room temperature. For

Imaging of Metals, Metalloids, and Non-metals

55

fresh tissues (e.g., plant tissues), the application of a cooled laser ablation chamber – commercially available from selected laser ablation companies or as developed in the author’s lab using two Peltier elements behind the target holder made of aluminum (38)– is recommended. The laser ablation chamber is coupled directly via a connection tube to the ICP torch of a quadrupole-based inductively coupled plasma mass spectrometer (ICP-QMS) with or without a collision cell, to a double-focusing sector field ICP-MS with single ion collection (ICP-SFMS), or to an multiple ion collector ICP-MS (MC-ICP-MS) (1). In LA-ICP-MS, different types of mass spectrometers with single ion collection are employed. For imaging ICP-MS measurements of tissues, quadrupole mass spectrometers are common. Figure 3.1 shows a schematic of an imaging mass spectrometer (LA-ICP-MS) for elemental mapping using a laser ablation system for the ablation of biological tissue in an inert atmosphere, an inductively plasma ion source and a quadrupole ion separation system. Similarly, the laser ablation system was coupled to double-focusing sector field ICP-MS or time-of-flight analyzers. LA-ICP-MS with sector fields offers the highest sensitivity achievable for the imaging of selected trace elements (for difficult to analyze elements such as Se) (27) or for spatially resolved measurements in the nanometer range (39–41). In our experiments with LA-ICP-MS, we used laser ablation systems (Nd:YAG at a wavelength of 266 or 213 nm from CETAC and from NewWave) coupled to ICP-QMS Agilent (with an octopole reaction cell), Elan 6100, or to double-focusing sector field ICP-SFMS (with reverse Nier–Johnson geometry) Element from Thermo Fisher Scientific. Time-of-flight (ToF) mass analyzers in LA-ICP-MS were applied for fast measurements of transient signals. Due to its relatively low sensitivity, LA-ICPToF-MS is restricted in its imaging of trace elements in biological tissues.

atmospheric pressure

laser beam

high vacuum

inductively coupled plasma

lens

ion detector Ar carrier gas flow

n

n

n

nn

+

n

load coil sample

laser ablation chamber

collision cell sampler

quadrupole analyzer

Fig. 3.1. Instrumental outline of LA-ICP-MS (laser ablation inductively coupled plasma mass spectrometer), the laser ablation system is coupled to an ICP quadrupole mass spectrometer with hexapole collision cell.

56

Sabine Becker and Susanne Becker

2.1. Measurement Procedure of Imaging LA-ICP-MS

Because soft materials’ biological tissues are easy to ablate from a glass substrate, an Nd:YAG laser with a wavelength of 266 nm is sufficient for complete ablation of sample. The optimum operating laser power density in all our experiments was found to be 1 × 109 W cm−2 . Mass spectrometric measurements by LA-ICP-MS for 2D imaging of biological tissues are performed by line scan (line per line) ablation of thin tissue sections with a focused laser (Nd:YAG) beam. The diameter of the laser crater can be varied during imaging mass spectrometric measurements of the thin tissue section between 50 and about 200 µm. The measurement time for imaging LA-ICP-MS of biological tissues (up to several hours) depends on the size of the tissue area analyzed or gel section selected and the laser scan speed applied (varies between 20 and 60 µm s−1 ). The spot size and laser scan speed were optimized to obtain images with high spatial resolution. In our routine measurements on rat/mouse brain tissues by LA-ICP-QMS, we applied a spatial resolution of 120 µm and a scan speed of 30 µm s−1 . In our experiments, the distance between the lines was only 10 µm, so that consequently most biological tissue was ablated. Due to the significantly higher sensitivity of LAICP-SFMS compared to LA-ICP-QMS the spatial resolution was reduced to 50 µm. A defined sample area (in general, several cm2 ) of a thin section of brain tissue (thickness: 20 µm) was scanned with a focused laser beam. Ion intensities of the analytes, e.g., of 64 Zn+ , were measured by LA-ICP-MS within the area of interest in different regions of the human brain tissue (e.g., human hippocampus). In order to compare measurements it is important that the experimental conditions are constant and the data obtained have high reproducibility. For the imaging of elemental distribution in protein spots blotted onto membranes of a conventional size, Jakubowski and coworkers (35) investigated different laser ablation cell geometries for the continuous ablation of nitrocellulose (NC) membranes during linear translation of the cell. The final cell has a volume of about 11.3 cm3 in which single-shot signals are washed out in less than a second so that the translation velocities of up to 1.5 mm s−1 can be applied to baseline separate structures with a distance of less than 2 mm, as is demonstrated using the example of a phosphoprotein mixture separated by SDS-PAGE and blotted onto a NC membrane. The whole procedure from the sample preparation of thin sections of biological tissue by cryo-cutting via the imaging procedure in LA-ICP-MS, including the scanning (line by line) of tissue in the selected brain area of mouse brain by a focused laser beam to measure ion intensities for analyte ions as a function of time and for a final evaluation of data in order to produce

Imaging of Metals, Metalloids, and Non-metals

Cryocutting

57

Thickness: 20 µm

Scanning of thin tissue slices using a focused laser beam (line per line)

Imaging of metals and non-metals by LA-ICP-MS

Focused laser beam 64Zn+

56Fe+

Data acquisition

Quantification using matrix matched laboratory standards

Quantitative images of metals

Fig. 3.2. Schematic of mass spectrometric imaging procedure by LA-ICP-MS on thin section of mouse brain tissue.

quantitative images of metals, is illustrated in Fig. 3.2. The metal images obtained, e.g., the Zn and Fe distributions, in mouse brain tissue (Parkinson’s disease) shown in this figure correlates with immunostained and autoradiographic images and photographs of the investigated slice. 2.2. Evaluation and Quantification of Analytical Data

No commercial software is yet available for the evaluation of analytical LA-ICP-MS data in order to obtain quantitative images of biological tissues. We developed different software packages with the aim of obtaining 3D metal or non-metal distributions, e.g., in rat brain, in order to correlate the data with MRI (magnet resonance images), immunostained, autoradiographic, and histochemical images of the same rat brain obtained at Institute of Medicine of Forschungszentrum Jülich, Germany. Furthermore, new quantification strategies were developed and adapted to the biological matrix investigated (21).

2.2.1. Quantification Using Synthetic Matrix-Matched Laboratory Standards

Quantitative images of elements can be obtained using prepared matrix-matched laboratory standards for calibration as demonstrated in several previous papers (1, 21, 42). The preparation protocol for synthetic matrix-matched laboratory standards for the calibration of imaging LA-ICP-MS for biological tissues is summarized in Fig. 3.3. In general, different synthetic laboratory standards with elements of interest in well-defined concentrations were prepared. For example, five slices of the same biological tissue were spiked with selected standard solutions at

58

Sabine Becker and Susanne Becker

Five laboratory synthetic standard solutions with elements of interest (Cu, Zn, Fe, Pb, Cd, U etc.) and well-defined concentrations were prepared

Five slices of the same biological tissue (each of about 0.65 g) were spiked with selected standard solutions (final concentration of Cu, Zn, Fe in brain tissue: 50, 20, 10, 5, 1 µg g–1 and of Pb, Cd and U: one order of magnitude lower) An additional slice was not spiked and was used for blank correction

Biological tissue were properly mixed and centrifuged at 5000 rpm for 5 min After that samples were frozen under the temperature –50 ºC

Frozen homogenized spiked biological tissues were cut with microtone in a thickness of 20 μm in the way similar to the sample done and placed onto the glass substrate

Prepared synthetic laboratory standards were further used for calibration of LA-ICP-MS measurements to determine concentration of selected elements inbiological tissues

Fig. 3.3. Preparation protocol of synthetic matrix-matched laboratory standards for calibration of LA-ICP-MS images.

defined concentration, mixed and frozen at a temperature of –50ºC, cut in thin slices, and placed onto the glass substrate. An additional unspiked slice was used for blank correction. The laboratory standards for the calibration of LA-ICP-MS images were measured together with biological tissues in the same measurement cycles, and consequently an ideal matrix matching was obtained. Matrix-matched homogenized laboratory standards with well-defined element concentrations of analytes were prepared and utilized for the quantification of LA-ICP-MS data in imaging brain tissues in routine mode in our laboratory. Matrix-matching standards were used to constitute calibration curves, where the regression coefficient of the calibration curves obtained was typically >0.9 for all analytes investigated. 2.2.2. Quantification Using Standard Reference Materials

Jackson et al. (43) proposed using pressed pellets of biologically certified reference materials for calibrations of rat brain images. The authors determined the elemental distributions and concentrations of Cu, Zn and Fe by LA-ICP-QMS (laser spot size: 60 µm and scan speed: 120 µm s−1 ) in whole rat brain sections (100 µm thickness) and the physiologically important elements P and S were also analyzed. The distributions and concentration ranges for these elements demonstrate the utility of this technique for rapid mapping of brain thin sections (43). Kidness et al. (29) described an analytical procedure for 2D mapping of copper and zinc in a section of sheep liver. For quantification purposes, the authors applied a homogeneous Certified Reference Material (CRM; LGC 7112, pig liver).

Imaging of Metals, Metalloids, and Non-metals

59

To quantify images of parts of heavy-metal-tolerant plants (leaves of Elsholtzia splendens) in order to study the hyperaccummulation of Cu and Zn, the standard reference material, NIST SRM 1515 apple leaves, was doped, mixed and homogenized with analytes of defined concentration. The quantification was performed via an external calibration procedure by measuring the calibration curves. Another standard reference material, NIST SRM 1547 Peach Leaves, was used to validate the analytical procedure (44, 45). 2.2.3. Solution-Based Calibration in LA-ICP-MS

3. Figures of Merit of LA-ICP-MS Imaging on Biological Tissues

Several solution-based calibration techniques (external calibration, standard addition and isotope dilution technique) developed in the author’s laboratory represent new calibration strategies for the quantification of analytical data in LA-ICP-MS (1). During the laser ablation of a thin section of biological tissue, defined standard solutions with increasing concentration were nebulized, whereby the calibration of the analytical method was performed by a standard addition mode (21). Online isotope dilution technique LA-ICP-MS is proposed as the method of choice for quantitative element mapping and imaging even if no reference materials are available (42). For online solution-based calibration with LA-ICP-MS, a microflow total consumption nebulizer DS-5 (CETAC Technologies, Omaha, NE, USA) was inserted into the laser ablation chamber (21). It is more convenient to use solution-based calibration because such calibration can be performed easily, quickly and in any concentration range for many elements. The application of the online isotope dilution technique in LA-ICP-MS using a microflow nebulizer inserted into the laser ablation chamber to biological standard reference materials (e.g., apple leaves) yielded accurate analyte concentrations (42).

As a powerful trace element analytical technique with multielement capability, LA-ICP-MS allows the production of a multitude of metals, metalloids and non-metal images quasisimultaneously. Ion images were evaluated on the basis of the imaging procedure by the newly developed LA-ICP-MS technique. In order to validate the ion images, at least two isotopes were analyzed if available (e.g., 63 Cu+ and 65 Cu+ or 64 Zn+ and 66 Zn+ ), and it is proposed that the elemental distribution in neighboring tissue slices should be performed under the same experimental parameters. Possible isobaric interferences of singlecharged analyte ions with polyatomic ions or double-charged

60

Sabine Becker and Susanne Becker

ions were studied carefully and can be generally minimized using double-focusing sector field ICP-MS at the required mass resolution or by applying collision-induced reactions in quadrupolebased ICP-MS with a collision/reaction cell. In general, fewer interference problems occur in the LA-ICP-MS analysis of thin tissues in dry plasma compared to ICP-MS measurements of aqueous solutions. The reproducibility of the imaging LA-ICP-MS method for the analysis of thin cross sections of biological tissues was studied by Zoriy and Becker (20). A roughly 3% reproducibility of the newly developed LA-ICP-MS procedure for homogeneous tissues (e.g., thin sections of matrix-matched laboratory (44) standards) was observed (see top right of Fig. 3.4). Dependences of LA-ICPMS signal intensities of 63 Cu+ on the laser beam diameter studied using synthetically prepared homogeneous brain laboratory standard doped with 10 µg g−1 of Cu is shown on the left in Fig. 3.4. In further studies, several adjacent inhomogeneous human brain sections of a thickness of 20 µm were cut from the same tissue and screened to compare the distribution of the elements of interest (e.g., Zn, Cu and C) and analyzed by LA-ICP-MS to discover how reproducible these profiles are within the neighboring sections. Figure 3.4 shows the quantitative Zn image of inhomogeneous human brain tissue and the results of reproducibility studies

3.1. Study of Reproducibility of Imaging LA-ICP-MS

63Cu+

1000000 900000 800000 700000 600000 500000 400000 300000 200000 100000 0

Laser beam diameter 110 µm 90 µm

+

70 µm 50 µm

0

500

1000

Cu ion intensity, cps

b)

35 µm

1500

63

Ion intensity, cps

a)

100000 90000 80000 70000 60000 50000 40000 30000 20000 10000 0

Reproducibility -2.3%

0

2000

1

2

3

4

5

6

7

8

9

10

line

Time, s

c)

d) Zn c

c

a

Concentration, ppm

Fig. 3.4. (a) Measured ion intensity of 63 Cu+ for different laser beam diameters and (b) reproducibility of imaging LAICP-MS on homogenized synthetic laboratory standard. A (c) Zn image of an inhomogeneous human brain tissue and (d) reproducibility of five adjacent slices of these brain tissue. (a) and (b) reprinted with permission from Elsevier.

Imaging of Metals, Metalloids, and Non-metals

61

for Zn imaging by LA-ICP-MS of five neighboring sections. The values obtained for the reproducibility in the three selected zones (marked in the left-hand figure) were in the range of 5.1–6.7%. Kidness et al. (29) characterized laser ablation inductively coupled plasma time-of-flight mass spectrometry (LA-ICP-MS) for imaging of biological tissues in terms of precision. Precisions for thin sections of sheep liver reached 27–59% (raster scan) and 9–47% (line scan) RSD for copper, whereas the precision for zinc was significantly better, 8–18% (raster scan) and 4–21% (line scan) RSD. 3.2. Detection Limits

4. Application Fields of Mass Spectrometric Imaging by LA-ICP-MS on Biological Tissues and Gels

The detection limits obtained for images measured by LAICP-MS depend on the experimental equipment and operating parameters employed and thus differ for different elements. The detection limits varied between microgram per gram and nanogram per gram. For example, the radionuclides Th and U were detected in biological tissues with highest sensitivity. Both radioactive metals were found to be homogeneously distributed at the ultratrace level in human hippocampus. The detection limits were determined for both elements in the low-nanogram per gram range. Jackson et al. (43) achieved sub-microgram per gram detection limits by imaging ICP-QMS on rat brain tissues (100 µm thickness) for the essential trace elements Cu and Zn. In the following section, selected examples of different applications of the LA-ICP-MS imaging technique will be discussed.

LA-ICP-MS enables imaging of thin biological tissue sections of plants and animals to study the distribution of essential and beneficial elements such as Zn, Cu, Fe, S, P, Se and Mn as well as of toxic and also radioactive metals (e.g., Hg, Pb, Cd, Th and U) with a spatial resolution in the micron range. In contrast to imaging by MALDI-MS, no application of matrix is required in LA-ICP-MS (nor in SIMS). A complete laser ablation of biological tissue is performed at higher laser energy in an atmospheric laser ablation chamber. The advantage of imaging LA-ICP-MS compared to SIMS, the latter operating in a high-vacuum ion source, is that no charging of the sample surface by the interaction of a focused laser beam with non-conducting material occurs and significantly fewer matrix effects are observed. These properties of LA-ICP-MS allow an easy quantification of the measured ion intensity. In the following section, the application of LA-ICP-MS for the imaging of metals, Se and non-metals will be discussed with

62

Sabine Becker and Susanne Becker

selected examples concerning brain samples, animal, or plant tissues. In addition, the new strategy of combining LA-ICP-MS imaging to detect metal-containing proteins and phosphoproteins and their elucidation by biomolecular mass spectrometry will be described. In Fig. 3.5 different application fields of mass spectrometry imaging by LA-ICP-MS are summarized. In the life sciences, in the author’s laboratory these applications focus mainly on the quantitative imaging of essential, beneficial, and toxic metals; metalloids and non-metals in thin tissue sections, e.g., in order to study diseased neurodegenerative tissue (Alzheimer’s and Parkinson’s disease) or tumor growth and stroke in the brain compared to normal tissue. In a second step, metalloproteins and phosphoproteins are observed after the gel electrophoretic separation of proteins (in 1D or 2D gels) by LA-ICP-MS, whereby MALDIMS and/or ESI-MS is employed for the identification of metalloproteins (phosphoproteins) in gels after spotting and tryptic digest. Nowadays, LA-ICP-MS imaging is employed to characterize the distribution of metal-containing proteins in gels (5). Alzheimer‘s

Normal brain functions

Parkinson‘s

Neurodegenerative diseases Brain tissues

Tumor

Epilepsia

growing therapy control

Schizophrenia Wilson‘s

Stroke

Brain aging

Huntington’s

Brain Research Metallomics

Mass Spectrometric Imaging Single cells Cell organells

LA-ICP-MS + MALDI/ESI-MS

of Metals, Metalloids and Nonmetals by LA-ICP-MS Metal-binding proteins

Biomonitoring Animals Kidney, liver, spleen, heart, lung, bones, teeth, hairs

bioavailability (e.g., Fe, Zn, Cu, Se, Ca)

Uptake

Plants Roots, leaves, stems

Transport Accumulation

toxicity (e.g., Cd, Pd, Hg, U)

Fig. 3.5. Application fields of mass spectrometric imaging by LA-ICP-MS in brain research and for biomonitoring of essential and toxic metals.

4.1. Multielemental Imaging by LA-ICP-MS on Brain Tissues

In the first example, the multielement capability of LA-ICPMS imaging will be demonstrated. The element distribution of selected essential metals in rat brain tissues was determined in the routine mode using a quadrupole ICP-MS (Agilent 7500 ce) with the laser UP 266 ablation system from NewWave with a lateral resolution in the 120 µm range. As illustrated in Fig. 3.6, the distribution of analyzed metals (ion images are shown here) such as

Imaging of Metals, Metalloids, and Non-metals

Fe

B

C

Zn

Mg

P

Cu

K

S

Na

Cl

63

Ion Intensity max.

Fig. 3.6. Images of 56 Fe+ , 64 Zn+ , 63 Cu+ , 11 B+ , 24 Mg+ , 39 K+ , 23 Na, 13 C+ , 31 P+ , 34 S+ , and 35 Cl+ ions of control rat brain measured by LA-ICP-QMS (Agilent 7500 ce and NewWave UP 266).

56 Fe+ , 64 Zn+ , 63 Cu+ , 11 B+ , 24 Mg+ , 39 K+

and

23 Na

and selected and is mostly different non-metals like (Becker et al. (2009), unpublished data). A layered structure was found for Fe and Zn in the rat brain cortex. However, the distribution of metal and non-metals also differs in the hippocampus. Whereas the highest Zn content was detected in the hippocampus (CA3 region), correlating with a slight enrichment of Mg, for other elements depletion in the hippocampus tended to be found. The phosphorus image measured by LA-ICP-MS using a quadrupole analyzer shows a similar distribution to that of carbon (see Fig. 3.6). The presented ion images of selected elements are sufficient to demonstrate the quite different element distribution in the brain section investigated and may yield insights on the role of these elements in selected brain regions. The quantitative Zn and Cu distribution in several sub-regions of the human brain hemisphere was determined by LA-ICP-SFMS. Synthetic matrixmatched laboratory standards were applied for quantification procedures of metal images (19). The distribution of Zn and Cu in a human brain hemisphere measured by LA-ICP-SFMS using a large laser ablation chamber from Cetac is illustrated in Fig. 3.7. As a result of LA-ICP-MS studies, we found a layered structure with higher Zn and Cu content in gray matter compared to white matter. The concentration profiles of Zn and Cu in the hippocampus area and in tissues from other regions of the human brain were investigated using LA-ICP-MS by Becker et al. (19, 21) It was found that copper and zinc are localized differently in the brain 13 C+ , 31 P+ , 34 S+

35 Cl+

64

Sabine Becker and Susanne Becker size, mm 100 Ion Intensity max.

50

Cu 0

Cu 75

150 size, mm

Zn

Fig. 3.7. Images of Cu and Zn in part of human brain and Cu distribution in whole hemisphere tissue measured by LA-ICP-SFMS using a large laser ablation chamber (Element and Cetac LSX 200).

regions analyzed with the highest concentration in the hippocampus. The distribution of toxic Pb was first measured in the human hippocampus in the 0.2 µg g−1 concentration range (19). 4.2. Study of Metal Distribution in Diseased Brain Region by LA-ICP-MS

The study of the distribution of metals in brain tissue is of special relevance in the tumor region compared to control tissue. It is well known that metals can also catalyze cytotoxic reactions leading to DNA modification (such as the Fenton reaction) or toxic processes at high concentrations that might be involved during tumor progression. Metal distribution in human brain tumor (human glioblastoma multiforme – GBM) and also tumor growth after injecting cultured tumor cells into rat brains were studied in tissue sections by imaging LA-ICP-MS in the author’s lab (31, 46, 47). A double-focusing sector field LA-ICP-SFMS (Element, Cetac LSX 200) was employed to measure the ion intensities of two essential elemental (Cu and Zn) and toxic metals (Pb and U) within the tumor area and the surrounding region invaded by GBM as well as in control tissue. The size of the brain samples was approximately 1 cm2 . The quantitative determination of copper, zinc, lead and uranium distribution in brain tissues by LA-ICP-MS was performed again using prepared matrix-matched laboratory standards doped with these elements of interest. An inhomogeneous

Imaging of Metals, Metalloids, and Non-metals

65

distribution was observed for all the metals measured. The limits of detection (LODs) obtained for Cu and Zn were 0.34 and 0.14 µg g−1 , respectively, while LODs of 12.5 and 6.9 ng g−1 were determined for Pb and U, respectively. A correlation was found between LA-ICP-MS and receptor-autoradiography results. The receptor-autoradiography technique using A1AR and pBR for labeling was employed. Regarding A1AR, we used the specific A1 adenosine receptor (A1AR)–ligand, 3HCPFPX [3 H-cyclopentyl-3-(3-fluoropropyl)-1-propylxanthine], which has been shown to specifically label the invasive zone around GBMs. The peripheral benzodiazepine receptor was labeled with 3 H-Pk11195 [3 H-1-(2-chlorophenyl)-N-methyl-N(1-methylpropyl)-3-isoquinoline-carboxamide]. Figure 3.8 illustrates the lateral distributions of Cu, Zn, Pb and U in thin sections of the glioblastoma human brain sample measured by LA-ICPSFMS together with the autoradiographic images: (a) peripheral benzodiazepine receptor (pBR) and (b) A1 adenosine receptor ligand (A1A).

a)

Cu Pb

Concentration, ppm Concentration, ppm

Zn

U

Concentration, ppm

Concentration, ppm

b)

c)

Fig. 3.8. (a) Cu and Pb distribution in glioblastoma multiforme sample measured by LA-ICP-MS, (b) peripheral benzodiazepine receptor (pBR) and (c) A1 adenosine receptor ligand (A1 A).

66

Sabine Becker and Susanne Becker

Metal concentrations were normalized for the cellular density, which was 1.3–4.8 times higher in tumors in comparison to control tissue. After normalization, the concentrations of copper and zinc were 3.9–9.1 and 2.4–9.6 times higher in the peritumoral zone compared to the solid tumor, respectively, whereas concentrations of lead and uranium were not elevated. Dehnhardt et al. (47) demonstrated that the increased copper and zinc levels were found in a zone which is characterized by a relatively high A1AR (adenosine receptor) density. A systematic study of several different rat brain samples infected with tumor cells (different incubation time) was performed by Zoriy et al. (46) The analyzed sections were quantitatively imaged for their P, S, Fe, Cu and Zn content. In addition, 13 C+ was monitored in all LA-ICP-MS measurements and its application as an internal standard was evaluated. The results of the measurements were compared with respect to the difference of the element content in control and tumor tissues. As an example, in Fig. 3.9 the lateral distribution images of Cu (b) and Zn (c) measured by LA-ICP-MS are presented. The small tumor at the bottom is made visible by a slight enrichment of Zn and a depletion of Cu (marked). After tumor growth (upper figures), the tumor region is characterized by a significantly changed metal distribution compared to the normal hemisphere (right). The locations of tumor region in the photographs and LA-ICP-MS images are in agreement, with the average concentrations of Cu and Zn in the tumor areas determined as 15 and 17 µg g−1 , respectively. The Cu and Zn contents were partly enriched or

a)

d)

b)

e)

c)

f)

Fig. 3.9. Images of Cu [(b) and (e)] and Zn [(c) and (f)] distribution in cross section of rat brain sample with two different tumor sizes compared to photograph of these slices [(a) and (d)] (reprinted with permission from Elsevier).

Imaging of Metals, Metalloids, and Non-metals

67

depleted in the tumor tissue (right side of each image) in comparison to the control (left side) as shown in Fig. 3.9. At present, we are studying the distribution of metals (via the measurement of 64 Zn+ , 63 Cu+ , 56 Fe+ , 49 Ti+ , 23 Na+ , 24 Mg+ and 39 K+ ) and non-metals (via 13 C+ , 31 P+ , 34 S+ and 35 Cl+ ) in thin tissue sections of rat brain (20 µm tissue thickness, maximum analyzed area ∼180 mm2 ) after a photoinduced cortical infarct (medical rat brain studies on photothrombosis were carried at Forschungszentrum Jülich by Langen et al. (48, 49)) using LAICP-QMS in order to construct 3D metal and non-metal images of the rat brain (Becker et al. (2009), unpublished data). Examples of ion images of selected metals (Fe, Zn, Cu, and Na) and non-metals (C, S, P, and Cl) measured on a rat brain section with a photoinduced infarct by imaging LA-ICP-MS are summarized in Fig. 3.10. The position of the measured slice in the rat brain with photothrombosis is marked in Fig. 3.10. This analytical technique was also employed to measure the very abundant alkali metals K and Na in control rat brain, as well as difficult-toionize halogens like chlorine. A strong correlation of Cl with the Na distribution in rat brain tissue was found. In the infarct region,

a)

Fe

C

b)

Photothrombosis

Zn

P

Cu

S

Na

Cl

analyzed slice

Ion Intensity max.

Fig. 3.10. (a) Distribution of Fe, Zn, Cu, Na, C, P, S and Cl in rat brain tissue with photoinduced stroke measured by LA-ICP-QMS (Agilent 7500 ce and NewWave UP 266) and (b) photograph of rat brain with photothrombosis (46).

68

Sabine Becker and Susanne Becker

selected elements (Fe, Zn, P, Mg and Ti) are clearly enriched, whereas K is depleted. These imaging mass spectrometric results of thin sections of rat brain tissues provide novel information on the distribution of elements in rat brain tissues compared to control samples and may also be an important tool for analyzing brain tumor metabolism and pathogenesis. Investigations of metal distribution in brain tissue from a mouse with a tumor treated with cisplatin were produced by scanning 14 µm thin sections of kidney tissue using sector field LAICP-SFMS as described by Zoriy et al. (50). Platinum was inhomogeneously distributed in the tissue with concentrations in the range of 10–25 µg g−1 where an enrichment in the tumor region was observed (50). As a microanalytical technique, LA-ICP-MS was applied in the author’s laboratory for evaluating the absorption of Pt by individuals undergoing cancer therapy using cisplatin. The concentration of Pt along a single strand of hair from a patient who had been treated with cisplatin as a cytostatic drug was monitored by LA-ICP-MS (51). For quantification, Pt-spiked hair strands were prepared and analyzed. By scanning the whole strand width by LA-ICP-MS small variations of Pt concentration along the hair strand can be observed. The maximum concentrations of Pt found along the hair strands in the measured transient signals were 26.9±5.3, 14.7±3.3, 20.9±3.9, and 26.1±3.8 µg g−1 , which corresponded to four treatments of cisplatin administered to the patient at 3 week intervals. The platinum distribution found in the analyzed hair may contribute to the optimization of cisplatin therapy. In many neurodegenerative diseases, abnormal metal deposition has been observed within the brain (e.g., in Alzheimer’s, Parkinson’s, or Wilson’s disease) (4, 52). We studied the iron content and metal distribution in Parkinson’s rat brain samples by LA-ICP-MS (53, 32). Numerous biochemical abnormalities have been discovered in the brain since MPTP (methylphenytetrahydropyridine) has been used as a neurotoxin-induced form of Parkinson’s disease by destruction of specific neurons in the substantia nigra (SN). The oxidative stress-induced neuronal damage of tissue is combined with a significant accumulation of iron in SN (54). In Fig. 3.11, the iron distribution compared to zinc and copper is illustrated in Parkinson’s mouse brain treated with MPTP. An increased Fe enrichment in SN is clearly seen, whereby we found that the Fe content is significantly higher compared to normal brain tissue (32). Zinc measured in the same slice is enriched in the hippocampus. In addition, poisoning by toxic heavy metals like Pb, Cd, or Hg, even in very low amounts, can cause permanent damage to the brain and nervous system (55). More details on the role of metals in the pathophysiology and pathogenesis of

Imaging of Metals, Metalloids, and Non-metals 64Zn+

63Cu+

69 56Fe+

Ion intensity

Fig. 3.11. Distribution of Zn, Cu and Fe in Parkinson’s mouse brain measured by LAICP-QMS (Agilent 7500 ce and NewWave UP 266).

neurodegenerative disorders are described in the book Neurodegenerative Diseases and Metal Ions edited by Sigel et al. (56). The specific toxicity of trace metals and compounds largely depends on their bioavailability in different organs or compartments of the organism considered. Imaging LA-ICP-MS with a spatial resolution in the 100 µm range was applied to study heavy metal distribution (neodymium and uranium) in brain tissues for toxicological screening (57). Rat brain postmortem tissues were stained in an aqueous solution of neodymium (metal concentration 100 µg g−1 ) for 3 h. The incubation of heavy metal in thin slices of brain tissue was observed by an imaging mass spectrometric LA-ICP-MS technique. The LA-ICP-SFMS images of neodymium (57) on stained rat brain tissue (thickness 30 µm) are shown in Fig. 3.12. Imaging LA-ICP-MS allows structures of interest to be identified and the relevant dose range to be estimated. Nd 5 µg g–1

0

Fig. 3.12. New staining technique in medicine: neodymium distribution in rat brain tissue after treatment in neodymium solution.

The determination of the metal concentration (excess or deficiency) compared to normal tissue, the binding to proteins and the quantitative distribution of metals in brain tissues is of the highest significance for the study and treatment of neurodegenerative diseases and is linked to the development of mass spectrometric techniques on biological complex systems. 4.3. Imaging of Metals Bound to Proteins in 2D Gels

Metallomics and metalloproteomics are emerging fields addressing the role, uptake, transport, and storage of trace metals essential for protein functions. The methodologies utilized in

70

Sabine Becker and Susanne Becker

metallomics and metalloproteomics to provide information on the identity, quantity, and function of metalloproteins are discussed here. As an elemental mass spectrometric technique, LAICP-MS has mostly been employed to identify the metal bound to a protein and MALDI/ESI-MS to elucidate the structure, dynamics, and function of a metal–protein complex. Other approaches include X-ray absorption and X-ray fluorescence spectroscopy and bioinformatics sequence analysis. X-ray absorption spectroscopy utilizing a synchrotron radiation source is a powerful tool for providing a direct analysis of metal bound to proteins and proteomic metal distribution in biological matrices. With the advent of genome sequencing, a large database of protein primary structures has been established, and specific tools have been developed to identify metalloproteins in the genome sequences (58). A challenging analytical strategy in life science studies consists of the combined application of element imaging mass spectrometric techniques like LA-ICP-IMS and biomolecular mass spectrometry such as ESI- or MALDI-MS to elucidate the structure and sequence of metal- or phosphorus-containing proteins. In Fig. 3.13, the combination of imaging mass spectrometry by LA-ICP-MS and proteome analysis is shown schematically with the example of one part of the human brain (hippocampus). This new analytical strategy starts by imaging thin sections of tissues to first detect metals, metalloids and non-metals in order to study neurodegenerative diseases or tumor growth via the quantitative (mostly abnormal) metal distribution of essential elements (e.g., Cu, Zn, Fe, Se) and also toxic metals (like Pb, Cd). In our

Excision of protein spot and digestion

Cryocutting

n

i te

Human brain

Quantitative LA-ICP-MS Imaging

is

es

se

or

h op

tr

o

Pr

n

io

at

r pa

lec

Combination of LA-ICP-MS & MALDI/ESI-MS

e el

2D

g

Screening for metals and P Detection of metal-containing & P-proteins by LA-ICP-MS

Identification and structure analysis by MALDI/ESI-MS

Fig. 3.13. Imaging LA-ICP-MS combined with proteome studies on selected part of brain tissue (hippocampus).

Imaging of Metals, Metalloids, and Non-metals

71

example (see Fig. 3.13), the hippocampus was quantitatively analyzed with respect to copper. The zinc distribution is shown and lead was detected at a low concentration level. In the second step, the proteins were separated from the selected region of interest by 1D or 2D gel electrophoresis. Metalloproteins and phosphoproteins were detected by LA-ICP-MS using a new powerful and sensitive screening technique. The protein spots containing phosphorus or metals were cut out from a second gel produced under exactly the same conditions, and after tryptic digestion the proteins were identified and the sequence determined by MALDI- or ESI-MS. For example, the binding of Cu and Zn on bovine serum proteins was studied using a tracer experiment by LA-ICP-MS measurements (33). 1D BN-PAGE gels with separated bovine serum proteins after gel electrophoresis were doped with an enriched isotope copper tracer (65 Cu) as a function of time (from 30 s up to 24 h). In several protein bands in the 1D gel before the tracer experiment metal ions (mostly Zn) were found by LA-ICPMS, but after experiments using enriched 65 Cu a fast exchange of Zn bonded on bovine serum albumin by copper was observed. This experimental finding demonstrates the formation of copperbinding proteins during tracer experiments in the 1D gel. Furthermore, proteins of bovine serum were separated by 2D blue native gel electrophoresis with native conditions in the first and the second dimension. The 2D BN-Page in the pI range of 4– 7 after colloidal Coomassie staining is illustrated on the righthand side of Fig. 3.14. The wide band in the middle of the gel was identified as bovine serum albumin by MALDI-TOF-MS. 34 S+ , 63 Cu+ and 64 Zn+ bonded to the protein were detected by b)

a) pI 4

7

c)

Fig. 3.14. (a) 2D BN-PAGE of bovine serum stained with colloidal Coomassie. (c) S, Cu and Zn bonded to the protein were detected by (b) LA-ICP-MS in five line scans from the (BSA) protein band.

72

Sabine Becker and Susanne Becker

LA-ICP-MS of bovine serum albumin in five line scans of 2D BN gel as is illustrated on the left in Fig. 3.14. A correlation of the maximum intensity of sulfur from protein with the maximum intensity of 63 Cu+ and 64 Zn+ was observed. 2D gel electrophoresis (2D GE) is well known as a powerful separation technique, which allows the separation of thousands of proteins from complex protein mixtures. The high-resolution capability of 2D GE can be obtained by independent protein separation steps in the first and second dimensions. The first dimension of gel electrophoresis is isoelectric focusing (IEF), whereby the individual proteins of a protein mixture move to their isoelectric point in a pH gradient. The proteins therefore lose their net charge and their electrophoretic mobility. In the second dimension, these proteins are separated orthogonally by SDS-PAGE according to their molecular weight (MW). The proteins separated in 2D gels are visualized, for example, by the silver staining technique. LA-ICP-MS was employed as a useful screening technique to detect metals and/or phosphorus in separated protein spots. In Fig. 3.15 one cut of a 2D SDS-PAGE gel from a human brain sample from the somatomotor cortex is shown. The

1800 ion intensity 1600 [cps]

P31 (MR)

31P+

11

1400 1200

8 2

1000 800 Blank

600

9

7 6

1

10

4

3

400

12

5

14

13

15

200 0

a)

0

50

100

150

200

250

time, s

300

MALDI- FTMS of spot 14

ion intensity 20000 [cps] 18000

7

63Cu+

16000 14000 12000 Blank

10000 8000 6000

8

2

4000

1

2000

4

3

5

6

9

11 10

12

13

15

14

0

b)

0

50

100

150

ion intensity 6000 [cps]

250

200

300

time, s

8

64Zn+

5000

2 Blank

4000 3000

9 6

1 2000

3

1000

4

10 11

7

15 12

5

13

14

0

c)

0

50

100

150

200

250

300

time, s

Fig. 3.15. Detection of P-, Zn- and Cu-containing proteins in human brain somatomotor cortex on a 2D gel (SDS-PAGE) LA-ICP-SFMS and identification of selected protein spot 14 by MALDI-FTICR-MS (34).

Imaging of Metals, Metalloids, and Non-metals

73

cut was scanned with LA-ICP-MS for phosphorus, copper and zinc. Phosphorus or metals was found in several of the analyzed gel spots. Also a MALDI-FTICR-MS spectrum with the identified peptides of spot 14 is shown in Fig. 3.15. A phosphopeptide was also identified in this spot. A BN-PAGE gel for protein separation can be used instead of IEF as the first dimension of the gel electrophoresis. This technique was used on mitochondria samples from baker’s yeast as shown on the left in Fig. 3.16. On this gel, different cuts were scanned for elements of interest like phosphorus and metals. The results from the LA-ICP-MS measurement of one cut can be seen in Fig. 3.16 on the right-hand side. For example, the protein in spot 59 was identified by a database search after MALDI-FTICRMS (see Fig. 3.16) as Aac2p (major mitochondrial ATP/ADP translocator).

54

2000

41

126

18 42 43

21

2500

53 55

56

44 22 5 6 7

8

23 26 27

9 10 11

32 33

24

57 45

25 46 60 61 47 128 28 29 30 31

58 127

59 62

125 48

34 12

Blank

35 13

1500 1000

59 127 127 59 61 61

Gel Blank

60 60

c)

46 62 128 62 125 125

500 0 300

90000 80000 70000 60000 50000 40000 30000 20000 10000 0

450

600

protein spots

56Fe+

127 61

128

59 60

300

36

64Zn+

protein spots gel blank

39 40

16 17

3000

51

gel blank

15 2 3 4 19 20

3500

52

50

38 1

b)

49

37

14

Ion intensity, cps

a)

62

125

450

600

5

Time, s

Spot 59

Fig. 3.16. Transient signals of 64 Zn+ - and 56 Fe+ -containing proteins in yeast mitochondria on a 2D gel (blue native/SDSPAGE) LA-ICP-SFMS and identification of selected protein spot 59 by MALDI-FTICR-MS (63).

74

Sabine Becker and Susanne Becker

Using this new analytical strategy, it would be possible to image the metal or phosphorus-containing protein spots in 2D gel in order to demonstrate their lateral distribution. The application of the imaging technique developed for tissues to the metal distribution in protein spots on gels allows additional valuable information to be obtained, for example, on the quality of the separation of metal- and phosphorus-containing proteins. This is demonstrated for Zn-binding proteins in a 2D BN-PAGE gel of rat kidney water extract. In Fig. 3.17, the lateral distribution of Zn on 2D BN-Page gel in several protein spots was measured by LA-ICP-QMS (using ICP-QMS Agilent 7500 ce and laser ablation system UP 266 from NewWave).

Fig. 3.17. (a) A 2D BN-PAGE gel (rat kidney water extract) (b) showing the detection of (b) Zn-containing proteins by imaging of the gel section using LA-ICP-SFMS (5).

4.4. Further Studies by Imaging LA-ICP-MS on Biological Tissues

Of special interest is the quantitative determination of selenium distribution in biological tissues. Over the past three decades, selenium has been intensively investigated as an antioxidant trace element (59). It is widely distributed throughout the body, but is particularly retained in the brain, even in the case of prolonged dietary selenium deficiency. Changes in selenium concentration in blood and brain have been reported in Alzheimer’s disease and brain tumors. The functions of selenium are believed to be carried out by selenoproteins, in which selenium is specifically incorporated as the amino acid, selenocysteine. Several selenoproteins are expressed in the brain and possess antioxidant activities and the ability to promote neuronal cell survival (59). An analytical imaging technique was developed to analyze thin brain sections by LA-ICP-SFMS. Due to the low selenium concentration in brain tissue, the technique was verified on slug tissues. For the quantification of selenium, again matrix-matched standard reference materials doped with selenium were prepared and analyzed. The selenium distribution is illustrated, together with the photograph of cross section of a slug, in Fig. 3.18. By LA-ICP-MS measurements of selenium in snail tissue a layered structure with a higher selenium concentration, e.g., in the skin and different organs was found. The selenium concentration in a 100 µm thin section of

Imaging of Metals, Metalloids, and Non-metals

a)

75

b) Se

–1

concentration, µg g

Fig. 3.18. (a) Selenium distribution in slug tissue measured by LA-ICP-SFMS (Element and Cetac LSX 200) and (b) a photograph of the tissue slice.

snail tissue was observed to be up to 25 µg g−1 . The detection limit for selenium was found to be 150 ng g−1 (27). LA-ICP-MS was employed to study biomonitoring of metal contamination in longitudinal tissue sections of the marine snail (60). The results of imaging of essential metals (Cu, Fe and, Zn), toxic metals (Pb, Cd, Hg, and As), radioactive metals (U and Th) and two halogens (I and Cl) in snail tissue measured by LAICP-MS compared to a photograph of the slice are shown in Fig. 3.19. The mass spectrometric analysis yielded an inhomogeneous distribution for all elements investigated. The detection limits for the distribution analysis of Cu, Zn, Cd, Hg and Pb measured by LA-ICP-MS were in the microgram per gram range. photograph

Cu

Fe

Zn

Pb

Cd

Hg

As

Th

U

I

CI

Fig. 3.19. Imaging of essential and toxic metals in snail tissue measured by LA-ICP-MS (Agilent 7500 ce and NewWave) compared to photograph of slice (left top).

Metal images in the leaves, shoots and roots of plant tissues (fixed onto glass slides) were produced by LA-ICP-MS (ICPQMS Elan 6100 from Perkin Elmer/Sciex and Cetac LSX 200) in the author’s laboratory (61). Thin sections (30 and 40 µm) of tobacco (Nicotiana tobaccum) plant tissues were analyzed by LA-ICP-QMS with respect to Mg, Mn, Fe, Cu and Zn together with the laboratory standards under the same experimental

76

Sabine Becker and Susanne Becker

conditions. Together with the nutrient metals, toxic metals like Cd and Pb were detected in the tobacco leaf investigated. For example, the maximum Cd concentration found in the tobacco leaves was 5 µg g−1 . Quantitative imaging of the selected elements by LA-ICP-QMS revealed their inhomogeneous distribution in leaves, shoots, and roots. Certain terrestrial plants known as “metal hyperaccumulators” can accumulate high concentrations of potentially toxic metallic elements (such as zinc, manganese, nickel, cobalt, copper, selenium, cadmium or arsenic) in their leaves and stems without suffering from any impairment of growth. High-resolution secondary ion mass spectrometry (NanoSIMS) was employed for studies of metal distribution on longitudinal sections of Alyssum lesbiacum leaves (62). Smart et al. observed high concentrations of nickel in the peripheral regions of the large unicellular stellate leaf hairs (trichomes) and in the epidermal cell layer. Electron probe microanalysis (EPMA) was used to provide independent confirmation of elemental distribution in the specimens, but the superior spatial resolution and high chemical sensitivity of the nanoSIMS technique provided a more detailed image of elemental distribution in these biological specimens at the cellular level. Whereas EMPA allowed a quantification of the metal concentration in the percentage and sub-percentage range, SIMS only yielded qualitative ion images of a small analysis area but at excellent lateral resolution in the 50 nm range. Quantitative imaging of essential nutrients in the plant leaves was conducted using LA-ICP-MS to study the accumulation and distribution of metals and non-metals by metaltolerant/hyperaccumulator plants, that means the uptake and transport of nutrients in plants which are able to hyperaccumulate toxic metals. After 65 Cu treatment, the leaves of the Cutolerant plant E. splendens were scanned directly with a focused laser beam in a laser ablation chamber to investigate the accumulation of Cu and other essential nutrients in the leaves. The ablated material was transported with argon as the carrier gas to a quadrupole-based ICP-MS (ICP-QMS), and the ion intensities of 65 Cu+ ,39 K+ , 55 Mn+ , 31 P+ and 11 B+ were measured by ICP-QMS. For quantification purposes, synthetic laboratory standards were prepared from standard reference material (NIST SRM 1515 Apple Leaves) spiked with defined concentrations of analytes and used for calibration. The standards were measured together with the leaf samples. The quantification procedure was validated by standard reference material SRM NIST 1547 Peach Leaves using one-point calibration. The resulting calibration curves from the prepared standards showed good linearity with the correlation coefficients (R2 >0.996). The distribution profiles of Cu, K, Mn, P and B in the leaves of E. splendens were quantified using these calibration curves. As shown in Fig. 3.20, the shape and structure

Imaging of Metals, Metalloids, and Non-metals

K

Fe

Mn

P

77

B

Concentration max.

Fig. 3.20. Quantitative images of nutrient elements K, Fe, Mn, P and B measured by LA-ICP-MS in the leaves of E. splendens after Cu treatment.

of the leaves was clearly given by LA-ICP-MS imaging of these elements. After treatment, the maximum Cu content was found in the veins near the petiole and at the bottom edge around the petiole of the newly formed leaves. In a further study, the ligand binding with Cu in the leaves will be studied by biomolecular mass spectrometry. Newly developed imaging techniques using LA-ICP-MS as an elemental analytical technique in combination with MALDIMS as a biomolecular mass spectrometric technique offer the capability to provide new information on analyzed tissue samples in order to understand and explain basic processes in the life sciences.

5. Future Developments of Imaging by LA-ICP-MS in the Nanometer Range

In order to improve the lateral resolution of LA-ICPMS down to the nanometer-scale range, near-field LA-ICP-MS NF-LA-ICPMS was created by Becker et al. (39–41). This technique uses the near-field enhancement effect at the tip of a thin silver needle in a laser beam (Nd:YAG laser, wavelength – 532 nm) on the sample surface. The thin silver needle was etched electrolytically in an electrochemical cell using a droplet of citric acid as the electrolyte. A robust needle etching procedure was established to produce the thin needles with a tip diameter in the hundreds of nanometer range. The “sample-to-tip” distance was controlled via the measurement of a tunnel current between the needle and the sample surface (39). For nanolocal analysis by NF-LA-ICP-MS on soft

78

Sabine Becker and Susanne Becker

matter (for example, on 2D gels and biological samples), a smallvolume transparent laser ablation chamber was constructed and coupled to a double-focusing sector field ICP mass spectrometer. A small amount of soft sample material was ablated at atmospheric pressure by a single laser shot in the near field of the silver tip in the defocused Nd:YAG laser beam. By single-shot analysis on 2D gels and biological surfaces doped with uranium, an enhancement of the ion intensities of the transient signals of up to factor 60 was observed compared to a background signal. Using the near-field effect in LA-ICP-MS, a nanolocal analysis will be possible on biological samples with nanometer-scale spatial resolution. The present experiments on near-field LA-ICP-MS open up a new and challenging path for future applications in imaging elements in the nanometer range in the life sciences, biology and medicine, e.g., for analyses of single cells, cell organelles, or biological structures in the nanometer range in order to detect disease, but also in materials science, nanotechnologies, and nanoelectronics.

6. Conclusions Imaging mass spectrometric techniques using LA-ICP-MS was developed for imaging the distribution of metals and non-metals in thin sections of biological tissues. The quantitative imaging analysis of essential and toxic elements in biological tissues allows studies to be performed of element distribution, transport processes, bioavailability, and possible contamination. The results of imaging mass spectrometry using LA-ICPMS in combination with biomolecular mass spectrometry provide novel information on the distribution of elements and element species in biological tissues and enable the identification of protein structures.

Acknowledgments The authors would like to thank A. Matusch (Philipps University of Marburg), M. Wagner (Goethe University Frankfurt, Germany) and D. Salber (Forschungszentrum Jülich, Germany) for providing biological tissues and M. Zoriy and A. Zimmermann (Forschungszentrum Jülich, Germany) and B. Wu (Zhejiang University, Hangzhou, China) for technical support with LA-ICP-MS measurements.

Imaging of Metals, Metalloids, and Non-metals

79

References 1. Becker, J. S. (2007)Inorganic Mass Spectrometry: Principles and Applications. John Wiley & Sons, New York, NY. 2. Becker, J. S. (2002) Determination of trace elements in small amounts of environmental samples by ICP-MS: a review. Can J Anal Sci Spectr, 47, 98–108 3. Pozebon, D., Dressler, V., Becker, J. S., Matusch, A., Zoriy, M., Becker, J. S. (2008) Biomonitoring of essential and toxic elements in small biological tissues by ICP-MS. J Anal At Spectrom, 23, 1281–1284. 4. Hutchinson, R. W., Cox, A. G., McLeod, C. W., Marshall, P. S., Harper, A., Dawson, E. L., Howlett, D. R. (2005) Imaging and spatial distribution of ß-amyloid peptide and metal ions in Alzheimer’s plaques by LAICP-MS. Anal Biochem, 346, 225. 5. Becker, J. S., Zoriy, M., Wu, B., Matusch, A., Becker, J. Su. (2008) Imaging of essential and toxic elements in biological tissues by LA-ICP-MS. J Anal At Spectrom, 23, 1275–1280. 6. Punshon, T., Jackson, B. P., Lanzirotti, A., Hopkins, W. A., Bertsch, P. M., Burger, J. (2005) Application of synchrotron Xray microbeam spectroscopy to the determination of metal distribution and speciation in biological tissues. Spectrosc Lett, 38, 343–363. 7. Carmona, A., Cloetens, P., Devès, G., Bohic, S., Ortega, R. (2008) Nano-imaging of trace metals by synchrotron X-ray fluorescence into dopaminergic single cells and neuritelike processes. J Anal At Spectrom, 23, 1083–1088. 8. Todd, P. J., Schaaf, T. G., Chaurand, P., Caprioli, R. M. (2001) Organic ion imaging of biological tissue with secondary ion mass spectrometry and matrix-assisted laser desorption/ionization. J Mass Spectrom, 36, 355. 9. Chandra, S. (2003) SIMS ion microscopy as a novel, practical tool for subcellular chemical imaging in cancer research. Appl Surf Sci, 203–204, 679–683. 10. Touboul, D., Halgand, F., Brunelle, A., Kersting, R., Tallarek, E., Hagenhoff, B., Laprevote, O. (2004) Tissue molecular ion imaging by gold cluster ion bombardment. Anal Chem, 76, 1550. 11. Broersen, A., van Liere, R., Altelaar, A. F. M., Heeren, R. M. A., McDonnell, L. A. (2008) Automated, feature-based image alignment for high-resolution imaging mass spectrometry of large biological samples. J Am Soc Mass Spectrom, 19, 823–832.

12. McDonnell, L. A., Piersma, S. R., Altelaar, A. F. M., Mize, T. H., Luxembourg, S. L., Verhaert, P. D. E. M., van Minnen, J., Heeren, R. M. A. (2005) Subcellular imaging mass spectrometry of brain tissue. J Mass Spectrom, 40, 160–168. 13. Brunelle, A., Laprevote, O. (2007) Recent advances in biological tissue imaging with time-of-flight secondary ion mass spectrometry: polyatomic ion sources, sample preparation, and applications. Curr Pharm Des, 13, 3335–3343. 14. Brunelle, A., Touboul, D., Laprévote, O. (2005) Biological tissue imaging with timeof-flight secondary ion mass spectrometry and cluster ion sources. J Mass Spectrom, 40, 985–999. 15. Heeren, M. A., Mc. Donnell, L. A., Amstalden, E., Luxaebourg, S. L., Altelaar, A. F. M., Piersma, S. R. (2006) Why don’t biologists use SIMS? A critical evaluation of imaging MS. Appl Surf Sci, 252, 6827–6835. 16. Chandra, S., Tjarks, W., Lorey, D. R., Barth, R. F. (2007) Quantitative subcellular imaging of boron compounds in individual mitotic and interphase human glioblastoma cells with imaging secondary ion mass spectrometry (SIMS). J Microsc, 229, 92–103. 17. Becker, J. S., Zoriy, M., Becker, J. Su., Dobrowolska, J., Matusch, A. (2007) Imaging mass spectrometry by laser ablation inductively coupled plasma mass spectrometry in biological tissues and proteomics. J Anal At Spectrom, 22, 736–744. 18. Becker, J. S., Becker, J. Su., Zoriy, M., Dobrowolska, J., Matusch, A. (2007) Imaging mass spectrometry in biological tissues by laser ablation inductively coupled plasma mass spectrometry. Eur J Mass Spectrom, 13, 1–6. 19. Dobrowolska, J., Dehnhardt, M., Matusch, A., Zoriy, M., Koscielniak, P., Zilles, K., Becker, J. S. (2008) Quantitative imaging of zinc, copper and lead in three distinct regions of the human brain by laser ablation inductively coupled plasma mass spectrometry. Talanta, 74, 717–723. 20. Becker, J. S., Zoriy, M., Matusch, A., Salber, D., Palm, C., Becker, J. Su. (2010) Bioimaging of metals by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). Mass Spectrom Rev, 29, 156–175. 21. Becker, J. S., Zoriy, M., Pickhardt, C., Palomero-Gallagher, N., Zilles, K. (2005) Imaging of copper, zinc, and other elements

80

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Sabine Becker and Susanne Becker in thin section of human brain samples (hippocampus) by laser ablation inductively coupled plasma mass spectrometry. Anal Chem, 77, 3208–3216. Caprioli, R. M., Farmer, T. B., Gile, J. (1997) Molecular imaging of biological samples: localization of peptides and proteins using MALDI-TOF MS. Anal Chem, 69, 4751–4760. Stoeckli, M., Chaurand, P., Hallahan, D. E., Caprioli, R. M. (2001) Imaging mass spectrometry: a new technology for the analysis of protein expression in mammalian tissues. Nat Med, 7, 493–496. Norris, J. L., Cornett, D. S., Mobley, J. A., Andersson, M., Seeley, E. H., Chaurand, P., Caprioli, R. M. (2007) Processing MALDI mass spectra to improve mass spectral direct tissue analysis. Int J Mass Spectrom, 260, 212–221. Chaurand, P., Cornett, D. S., Caprioli, R. M. (2006) Molecular imaging of thin mammalian tissue sections by mass spectrometry. Curr Opin Biotech, 17, 431–436. Pohl, H. R., Abadin, H. G., Risher, J. F. (2006) Neurotoxicity of cadmium, lead and mercury. Metal Ions in Life Sciences (Sigel, A., Sigel, H., Sigel, R. K. O. eds.), 395–425, Wiley & Sons, New York, NY. Becker, J. S., Matusch, A., Depboylu, C., Dobrowolska, J., Zoriy, M. (2007) Quantitative imaging of selenium, copper, and zinc in thin sections of biological tissues (slugs-genus arion) measured by laser ablation inductively coupled of plasma mass spectrometry. Anal Chem, 79, 6074–6080. Feldmann, J., Kindness, A., Ek, P. (2002) Laser ablation of soft tissue using a cryogenically cooled ablation cell. J Anal At Spectrom, 17, 813–818. Kindness, A., Sekaran, N., Feldmann, J. (2003) Two-dimensional mapping of copper and zinc in liver sections by laser ablationinductively coupled plasma mass spectrometry. Clin Chem, 49, 1916–1923. Ghazi, A. M., Wataha, J. C., O′ Dell, N. L., Singh, B. B., Simmons, R., Shuttleworth, S. (2002) Quantitative concentration profiling of nickel in tissues around metal implants: a new biomedical application of laser ablation sector field ICP-MS. J Anal At Spectrom, 17, 1295. Becker, J. S., Zoriy, M., Becker, J. Su., Pickhardt, C., Damoc, E., Juhacz, G., Palkovits, M., Przybylski, M. (2005) Determination of phosphorus copper and zinc containing human brain proteins by LA-ICP-MS and MALDI-FTICR-MS. Anal Chem, 77, 5851–5860.

32. Matusch, A., Depboylu, C., Palm, C., Wu, B., Höglinger, G. U., Schäfer, M. K.-H., Becker, J. S. (2010) Cerebral bio-imaging of Cu, Fe, Zn and Mn in the MPTP mouse model of Parkinson’s disease using laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). J Am Soc Mass Spectrom, 21, 161–171. 33. Becker, J. Su., Pozebon, D., Dressler, V., Lobinski, R., Becker, J. S. (2008) LA-ICPMS studies of zinc exchange by copper in bovine serum albumin using an isotopic enriched copper tracer. J Anal At Spectrom, 23, 1076–1082. 34. Becker, J. Su., Mounicou, S., Zoriy, M. V., Becker, J. S., Lobinski, R. (2008) Analysis of metal-binding proteins separated by non-denaturating gel electrophoresis using matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). Talanta, 76, 1183–1188. 35. Feldmann, I., Koehler, C. U., Roos, P. H., Jakubowski, N. (2006) Optimisation of a laser ablation cell for detection of heteroelements in proteins blotted onto membranes by use of inductively coupled plasma mass spectrometry. J Anal At Spectrom, 21, 1006–1015. 36. Becker, J. Su., Zoriy, M., Przybylski, M., Becker, J. S. (2007) Study of formation of Cu- and Zn-containing tau protein using isotopic-enriched tracers by LA-ICP-MS and MALDI-FTICR-MS. J Anal At Spectrom, 22, 63–68. 37. Becker, J. S., Zoriy, M., Dehnhardt, M., Pickhardt, C., Zilles, K. (2005) Copper, zinc, phosphorus and sulfur distribution in thin section of rat brain tissues measured by laser ablation inductively coupled plasma mass spectrometry: possibility for small-size tumor analysis. J Anal At Spectrom, 20, 912–917. 38. Zoriy, M., Kayser, M., Izmer, A., Pickhardt, C., Becker J. S. (2005) Determination of uranium isotopic ratios in biological samples using laser ablation inductively coupled plasma double focusing sector field mass spectrometry with cooled ablation chamber. Int J Mass Spectrom, 242, 297–302. 39. Zoriy, M., Kayser, M., Becker, J. S. (2008) Possibility of nano-local element analysis by near-field laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS): new experimental arrangement and first application. Int J Mass Spectrom, 273, 151–155.

Imaging of Metals, Metalloids, and Non-metals 40. Zoriy, M., Becker, J. S. (2009) Near-field laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS): a novel elemental analytical technique at nanometer scale. Rapid Commun Mass Spectrom, 23, 23–30. 41. Becker, J. S., Gorbunoff, A., Zoriy, M., Izmer, A., Kayser, M. (2006) Evidence of near-field laser ablation inductively coupled plasma mass spectrometry (NF-LAICP-MS) at nanometre scale for elemental and isotopic analysis on gels and biological samples. J Anal At Spectrom, 21, 19–25. 42. Pickhardt, C., Izmer, A., Zoriy, M., Schaumloffel, D., Becker, J. S. (2006) Online isotope dilution in laser ablation inductively coupled plasma mass spectrometry using a microflow nebulizer inserted in the laser ablation chamber. Int J Mass Spectrom, 248, 136–141. 43. Jackson, B., Harper, S., Smith, L., Flinn, J. (2006) Elemental mapping and quantitative analysis of Cu, Zn, and Fe in rat brain sections by laser ablation ICP-MS. Anal Bioanal Chem, 384, 1618. 44. Zoriy, M., Becker, J. S. (2007) Imaging of elements in thin cross sections of human brain samples by LA-ICP-MS: A study on reproducibility. Int J Mass Spectrom, 264, 175–180. 45. Wu, B., Zoriy, M., Chen, Y., Becker, J. S. (2009) Imaging of nutrient elements in the leaves of Elsholtzia splendens by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). Talanta, 78, 132–137. 46. Zoriy, M., Dehnhardt, M., Matusch, A., Becker, J. S. (2008) Comparative imaging of P, S, Fe, Cu, Zn and C in thin sections of rat brain tumor as well as control tissues by laser ablation inductively coupled plasma mass spectrometry. Spectrochim Acta B, 63, 375–382. 47. Dehnhardt, M., Zoriy, M., Khan, Z., Reifenberger, G., Ekstrom, T. J., Becker, J. S., Zilles, K., Bauer, A. (2008) Element distribution is altered in a zone surrounding human glioblastoma multiforme. J Trace Elem Med Biol, 22, 17–23. 48. Langen, K. J., Salber, D., Hamacher, H., Stoffels, G., Reifenberger, G., Pauleit, D., Coenen, H., Zilles, K. (2007) Detection of secondary thalamic degeneration after cortical infarction using cis-4-18F-fluoro-Dproline. J Nucl Med, 48, 1482–1491. 49. Langen, K. J., Hamacher, K., Weckesser, M., Floeth, F., Stoffels, G., Bauer, D., Coenen, H. H., Pauleit, D. (2006) O-(2-

50.

51.

52.

53.

54.

55.

56. 57.

58. 59. 60.

81

[18F]fluoroethyl)-L-tyrosine: uptake mechanisms and clinical applications.Nucl Med Biol, 33, 287–294. Zoriy, M., Matusch, A., Spruss, T., Becker, J. S. (2007) Laser ablation inductively coupled plasma mass spectrometry for imaging of copper, zinc, and platinum in thin sections of a kidney from a mouse treated with cis-platin. Int J Mass Spectrom, 260, 102–106. Pozebon, D., Dressler, V., Matusch, A., Becker, J. S. (2008) Monitoring of platinum in a single hair by laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) after cisplatin. J Mass Spectrom, 272, 57–62. Qureshi, G. A., Syed, S. A., Parvez, S. H. (2007) Role of selenium, iron, copper and zinc. Oxidative Stress and Neurodegenerative Disorders, 719, Elsevier, Maryland Heights, MO. Depboylu, C., Matusch, A., Tibl, F., Zoriy, M., Michel, P., Riederer, P., Gerlach, M., Becker, J. S., Örtel, W. H., Höglinger, G. U. (2007) Glia protects neurons against extracellular human neuromelanin. Neurodegener Dis, 4, 218–226. Gerlach, M., Double, K. L., Götz, M. F., Youdim, M. B. H., Riederer, P. (2006) The role of iron in the pathogenesis of Parkinson’s disease. Neurodegenerative Diseases and Metal Ions (Sigel, A., Sigel, H., Sigel, R. K. O. eds.), John Wiley & Sons, New York, NY. Faa, G., Lisci, M., Caria, M. P., Ambu, R., Sciot, R. N., Nurchi, V. M.„ Silvagni, R., Diaz, A, Crisponi, G. (2001) Brain copper, iron, magnesium, zinc, calcium, sulfur and phosphorus storage in Wilson’s disease. J Trace Elem Med Biol, 15, 155–160. Sigel, A., Sigel, H., Sigel, R. K. O. (2006) Neurodegenerative Diseases and Metal Ions. John Wiley & Sons, New York, NY. Becker, J. S., Dobrowolska, J., Zoriy, M., Matusch, A. (2008) Imaging of uranium on rat brain sections using LA-ICP-MS: a new tool for the study of critical substructures affined to heavy metals in tissues. Rapid Commun Mass Spectrom, 22, 2768–2772. Shi, W., Chance, M. R. (2008) Metallomics and metalloproteomics. Cell Mol Life Sci (CMLS), 65, 3040–3048. Chen, J., Berry, M. J. (2003) Selenium and selenoproteins in the brain and brain diseases. J Neurochem, 86, 1–12. Santos, M. C., Wagner, M., Wu, B., Oehlmann, J., Cadore, S., Becker, J. S. (2009) Biomonitoring of metal contamination in a marine prosobranch

82

Sabine Becker and Susanne Becker

snail (Nassarius reticulatus) by imaging laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). Talanta, 80, 428–433. 61. Becker, J. S., Dietrich, R. C., Matusch, A., Pozebon, D., Dressler, V. L. (2008) Quantitative images of metals in plant tissues measured by laser ablation inductively coupled plasma mass spectrometry. Spectrochim Acta, B63, 1248–1252. 62. Smart, K. E., Kilburn, M. R., Salter, C. J., Smith, J. A. C., Grovenor, C. R. M. (2007) NanoSIMS and EPMA analysis of nickel

localisation in leaves of the hyperaccumulator plant Alyssum lesbiacum. Int J Mass Spectrom, 260, 107–114. 63. Becker, J. S., Zoriy, M., Krause-Buchholz, U., Becker, J. S., Pickhardt, C., Przybylski, M., Pompe, W., Roedel, G. (2004) In-gel screening of phosphorus and copper, zinc and iron in proteins of yeast mitochondria by LA-ICP-MS and identification of phosphorylated protein structures by MALDI-FT-ICRMS after separation with two-dimensional gel electrophoresis. J Anal At Spectrom, 19, 1236–1243.

Part II Protocols for MS Imaging of Distribution of Small Molecules Including Metabolites and Pharmaceuticals

Chapter 4 Lipid Detection, Identification, and Imaging Single Cells with SIMS Michael L. Heien, Paul D. Piehowski, Nicholas Winograd, and Andrew G. Ewing Abstract Time-of-flight secondary ion mass spectrometry (ToF-SIMS) can be utilized to map the distribution of various molecules on a surface with submicrometer resolution. Many of its biological applications have been in the study of membrane lipids, such as phospholipids and cholesterol. For these studies, the effectiveness of chemical mapping is limited by low signal intensity from various biomolecules. Because of the high-energy nature of the SIMS ionization process, many molecules are identified by detection of characteristic fragments. Cluster ion sources are able to increase ionization, leading to increased information collected from a surface. In this chapter, we highlight the utility of SIMS to image lipids at single cells. Particularly, we will describe sample preparation, data collection, and the analysis of lipids for two systems; rat oligodendrocytes and Tetrahymena thermophila. SIMS spectra yield information regarding lipid identity and concentration across cell surface. Key words: Time-of-flight (ToF), secondary ion mass spectrometry (SIMS), freeze-fracture, freezeetch, imaging, mass spectrometry, lipids, single cell.

1. Introduction Mass spectrometry imaging with time-of-flight secondary ion mass spectrometry (ToF-SIMS) has revealed the spatial distribution of chemicals on a surface (1). When applied to biological samples, this method offers spatial information on biologically relevant small molecules ( 8 cm) was recommended to produce dryer DHB particles, which can minimize the analyte migration induced by the solvent and yield much greater signal response (15). 5. The accelerator voltage of the SEM system was typically 5 kV. No carbon or gold was further coated to the matrix surface prior to the SEM analysis. 6. The formation of matrix crystals on the sample surface can greatly affect the results of imaging MALDI-MS. In Fig. 7.2, it is clearly shown that the size and surface distribution of DHB crystals are dramatically different when using various matrix deposition protocols. SEM characterization of matrix crystals can be used to assist the optimization of matrix coating and improve the quality of the MALDI images. 7. It was observed that a matrix thickness between 5 and 50 µm was sufficient for reasonable signal-to-noise ratio (s/n) of lipid ions in mouse brain tissue and other samples. 8. To avoid a clogging problem with the OCN, it is suggested to pump the matrix solution for 2–3 min before turning on the nebulizing gas. If the flow of matrix solution needs to be stopped because of changing the syringe, adding more matrix solution, or switching to nanopure water when the matrix application is finished, turn off the nebulizing gas first and keep pumping the matrix solution for another 2–3 min after there is no gas flow through the OCN. 9. The mass spectra and resulting ion images can be obtained with only 8–10 laser shots per spot in either the positive or the negative ionization mode. This suggests that the OCN matrix deposition system is able to generate a good matrix–analyte interaction, which promotes efficient laser desorption and subsequent ionization of lipids. Since the OCN system can yield mass spectra with high s/n ions via fewer laser shots, the data acquisition time was also greatly reduced, which is a considerable benefit for timeconsuming imaging mass spectrometry experiments. 10. The step size should be carefully selected. Although decreasing the step size may provide more details about the sample, it also causes longer data acquisition time,

144

Chen et al.

larger data files, and the possibility of matrix subliming off the sample in ultra-high vacuum (UHV) chamber, especially for samples prepared using dryer matrix coating conditions. 11. MALDI-MS spectra acquired from the hexb−/− mouse brain slices prepared by OCN matrix coating system showed several prominent ions of m/z 888.6, 1,132 and 1,383 (Fig. 7.3) localized in different regions of the brain. 12. Sphingolipids (sulfatide, ganglioside GM2, and asialoGM2 (GA2)) were distinctly visible in hexb−/− mouse brain samples by using OCN for matrix application. These ion images clearly demonstrate that the OCN system is useful for sample preparation for imaging MALDI-MS of lipids. The spatial distribution of sulfide subspecies ST d18:1/C22:0 (m/z 862.6), ST(OH) d18:1/h22:0 (m/z 878.6), ST d18:1/C24:1 (m/z 888.6), ST d18:1/C24:0 (m/z 890.6), ST(OH) d18:1/h24:0 (m/z 906.6), ST(OH) d18:0/h24:0 (m/z 908.6), and an unknown ion (m/z 868.6) displayed a remarkably similar pattern to the myelinated fiber (white matter) region of the cerebellum and a relatively even distribution in brain stem (c.f., H&E staining) (Fig. 7.4a–g). The localizations of potassiated ganglioside GA2 (d18:1/C18:0) (m/z 1,132), potassiated ganglioside GA2 (d20:1/C18:0) (m/z 1,160), ganglioside GM2 (d18:1/C18:0) (m/z 1,383), and ganglioside GM2 (d20:1/C18:0) (m/z 1,411), which are also known to accumulate in mice with this genetic defect (16), closely matched the granular cell region in cerebellum and produced no detectable ions in the brain stem region (Fig. 7.4 h–k). The imaging MALDI-MS results in Fig. 7.4 illustrate that various subcategories of sphingolipids are localized to specific regions of the brain. Therefore, this technology is a valuable complement to other types of “lipidomic” analysis, which uses homogenized extracts of the entire tissue which may miss potentially important regional changes in both the types and the amounts of the lipids present. 13. ESI-MS/MS analysis of the lipid extracts from the mouse brains was performed to confirm the structure of sulfatide, GM2, and GA2. For example, in negative ion mode, MS/MS of m/z 1,410.9 generates five major fragment ions corresponding to losses of different sugar moieties in the head group (Fig. 7.5a). The product ions of m/z 1,119.8, 916.8, 754.8, and 592.6 correspond to the Y-type glycosidic bond cleavage involving loss of NeuAc, NeuAc/GalNac, NeuAc/GalNac/Gal, and NeuAc/GalNac/Gal/Glc, respectively. The m/z 290.1

Imaging MALDI Mass Spectrometry

145

ions were produced by C-type cleavage and charge retention on the sialic acid with subsequent dehydration, which confirms the existence of a sialic acid moiety. 14. An MS3 experiment was performed on the Y0 fragment ion of m/z 592.6 to determine the ceramide backbone of the m/z 1,410.9 ion. The resulting MS3 spectra (Fig. 7.5b) showed secondary fragment ions of m/z 324, 308, 282, and 283, corresponding to S, T, U, and V + 16 fragments, respectively (17), revealing that the amide-linked fatty acid is stearate (C18:0). The ions of m/z 265 and 291 correspond to complimentary P and Q fragments, respectively (17), showing that the sphingoid base backbone is d20:1. Thus, this major species in hexb−/− mouse brain is ganglioside GM2 (d20:1/C18:0).

Acknowledgments The authors would like to thank Drs. Richard Browner and Facundo Fernandez for providing the OCN sprayer, Dr. Markus Stoeckli for sharing the modified MMIST software, and Lan Sun for SEM analysis. This work is supported by NIH GM069338 (Lipid MAPS) and seed funding from Georgia Institute of Technology for the Mass Spectrometry Bio-Imaging Center. References 1. Caprioli, R. M., Farmer, T. B., Gile, J. (1997) Molecular imaging of biological samples: localization of peptides and proteins using MALDI-TOF MS. Anal Chem, 69, 4751–4760. 2. Chaurand, P., Schwartz, S. A., Caprioli, R. M. (2004) Profiling and imaging proteins in tissue sections by MS. Anal Chem, 76, 86A–93A. 3. Rubakhin, S. S., Greenough, W. T., Sweedler, J. V. (2003) Spatial profiling with MALDI MS: distribution of neuropeptides within single neurons. Anal Chem, 75, 5374–5380. 4. McDonnell, L. A., Heeren, R. M. A. (2007) Imaging mass spectrometry. Mass Spectrom Rev, 26, 606–643. 5. Wang, L., May, S. W., Browner, R. F. (1996) Low-flow interface for liquid chromatography-inductively coupled plasma mass spectrometry speciation using an oscillating capillary nebulizer. J Anal Atomic Spectrom, 11, 1137–1146.

6. Reyderman, L., Stavchansky, S. (1996) Novel methods of microparticulate production: application to drug delivery. Pharm Dev Technol, 1, 223–229. 7. Perez, J., Petzold, C. J., Watkins, M. A., Vaughn, W. E., Kenttamaa, H. I. (1999) Laser desorption in transmission geometry inside a Fourier-transform ion cyclotron resonance mass spectrometer. J Am Soc Mass Spectrom, 10, 1105–1110. 8. Lake, D. A., Johnson, M. V., McEwen, C. N., Larsen, B. S. (2000) Sample preparation for high throughput accurate mass analysis by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Commun Mass Spectrom, 14, 1008–1013. 9. Basile, F., Kassalainen, G. E., Williams, S. K. R. (2005) Interface for direct and continuous sample-matrix deposition onto a MALDI probe for polymer analysis by thermal field flow fractionation and off-line MALDI-MS. Anal Chem, 77, 3008–3012.

146

Chen et al.

10. Fung, K. Y. C., Askovic, S., Basile, F., Duncan, M. W. (2004) A simple and inexpensive approach to interfacing highperformance liquid chromatography and matrix-assisted laser desorption/ionizationtime of flight-mass spectrometry. Proteomics, 4, 3121–3127. 11. Kirlew, P. W., Caruso, J. A. (1998) Investigation of a modified oscillating capillary nebulizer design as an interface for CE-ICP-MS. Appl Spectrosc, 52, 770–772. 12. Chen, Y., Allegood, J., Liu, Y., Wang, E., Cachon-Gonzalez, B., Cox, T. M., Merrill, A. H., Jr., Sullards, M. C. (2008) Imaging MALDI mass spectrometry using an oscillating capillary nebulizer matrix coating system and its application to analysis of lipids in brain from a mouse model of Tay-Sachs/Sandhoff disease. Anal Chem, 80, 2780–2788. 13. Cachon-Gonzalez, M. B., Wang, S. Z., Lynch, A., Ziegler, R., Cheng, S. H., Cox, T. M. (2006) Effective gene therapy in an authentic model of Tay-Sachs-related

14.

15.

16.

17.

diseases. Proc Natl Acad Sci U S A, 103, 10373–10378. van Echten-Deckert, G. (2000) Sphingolipid Metabolism and Cell Signaling, Pt B, Vol. 312, Academic Press Inc, San Diego, CA, 64–79. Hankin, J. A., Barkley, R. M., Murphy, R. C. (2007) Sublimation as a method of matrix application for mass spectrometric imaging. J Am Soc Mass Spectrom, 18, 1646–1652. Conzelmann, E., Sandhoff, K. (1978) Abvariant of infantile Gm2 gangliosidosis - deficiency of a factor necessary for stimulation of hexosaminidase a-catalyzed degradation of ganglioside Gm2 and glycolipid Ga2. Proc Natl Acad Sci U S A, 75, 3979–3983. Merrill, A. H., Sullards, M. C., Allegood, J. C., Kelly, S., Wang, E. (2005) Sphingolipidomics: high-throughput, structurespecific, and quantitative analysis of sphingolipids by liquid chromatography tandem mass spectrometry. Methods, 36, 207–224.

Chapter 8 Mapping Pharmaceuticals in Rat Brain Sections Using MALDI Imaging Mass Spectrometry Yunsheng Hsieh, Fangbiao Li, and Walter A. Korfmacher Abstract Matrix-assisted laser desorption/ionization-tandem mass spectrometric method (MALDI-MS/MS) has proven to be a reliable tool for direct measurement of the disposition of small molecules in animal tissue sections. As example, MALDI-MS/MS imaging system was employed for visualizing the spatial distribution of astemizole and its primary metabolite in rat brain tissues. Astemizole is a second-generation antihistamine, a block peripheral H1 receptor, which was introduced to provide comparable therapeutic benefit but was withdrawn in most countries due to toxicity risks. Astemizole was observed to be heterogeneously distributed to most parts of brain tissue slices including cortex, hippocampus, hypothalamic, thalamus, and ventricle regions while its major metabolite, desmethylastemizole, was only found around ventricle sites. We have shown that astemizole alone is likely to be responsible for the central nervous system (CNS) side effects when its exposures became elevated. Key words: MALDI, tandem mass spectrometry, astemizole, rat brain tissues, drug localization.

1. Introduction First-generation antihistamines provide symptomatic relief from allergies and the common cold to patients. However, their therapeutic potential is hampered by the sedation caused by their effects on histamine receptors in the brain (1–3). Secondgeneration antihistamine (astemizole, as an example) block peripheral H1 receptors were introduced to provide comparable therapeutic benefit without the CNS side effects under manufactures’ recommended doses (1–4). It was reported that astemizole was found to cause arrhythmias when drug exposures became S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_8, © Springer Science+Business Media, LLC 2010

147

148

Hsieh, Li, and Korfmacher

elevated and the cardiac toxicity was mainly due to the parent drug (5). As the site of drug administration is often distinct from the site of the drug target, in vivo drug distribution is a critical aspect of toxicological and pharmacological action. Autoradiography (6–8) has been a common method for measuring drug distribution in animal tissue sections. However, the major disadvantage in this radioautographic technique is that it detects total drugrelated materials but does not distinguish between the administered drug and its metabolites. This is because the measured radioactivity signals come from any drug-related material including the administered compound and its metabolites, which may confound data interpretation. In this work, the uptake and the retention of astemizole into rat brains with and without perfusion with saline solution using MALDI-MS/MS imaging technique (9–19) was investigated in order to resolve the possible causes of its CNS side effects.

2. Materials 2.1. Chemical Reagents

1. 2, 5-Dihydroxybenzoic acid (DHB) used as matrix substances for MALDI (Aldrich, Milwaukee, WI, USA) is dissolved in 50:50 acetonitrile in water as a 20 mg/ml matrix solution. 2. Astemizole used as an analyte (Sigma, St. Louis, MO, USA) is dissolved in acetonitrile at a concentration of 1 mg/ml and then serially diluted with 50:50 acetonitrile:water to the appropriate concentrations or is formulated with 0.4% methylcellulose dosing solution at a concentration of 10 mg/ml. 3. Acetonitrile (HPLC grade) (Fisher Scientific, Pittsburg, PA, USA) is used as organic solvent. 4. Methylcellulose (The Dow Chemicals, Piscataway, NJ, USA) is dissolved in water as 0.4% dosing solution. 5. Dulbecco’s Phosphate-Buffered Saline (D-PBS) (Invitrogen, California, CA, USA) for perfusion experiment.

2.2. Equipment

1. Water purification system (Millipore, Billerica, MA, USA). 2. Multistep pipette with 1 µl pipette tips (Eppendorf, Westbury, NY, USA). 3. A CM3050 cryostat (Leica Microsystems Inc., Bannockburn, IL, USA) designed for rapid freezing and sectioning of tissue samples was used to produce the high-quality frozen sections.

Mapping Pharmaceuticals in Rat Brain Sections

149

4. A Nikon microscope (Micron Optics, Cedar Knolls, NJ, USA) was used to record optical images to define the outline of tissue sections. 5. A glass reagent sprayer (Kontes Glass Company, Vineland, NJ, USA) was used to coat the matrix solution over the tissue sections. 6. A vacuum desiccator (Fisher Scientific, Pittsburg, PA, USA). 7. Opti-TOFTM MALDI plate system (Applied Biosystems, Foster City, CA, USA) provides hydrophobically coated MALDI sample plates (5×5 cm) for spotting and tissue imaging applications which include a reusable holder and insert and are compatible with all Applied Biosystems MALDI plateforms. 8. A QStar Pulsar (Applied Biosystems, Foster City, CA, USA) hybrid QqTOF mass spectrometer equipped with an oMALDITM (orthogonal MALDI) ion source and a nitrogen laser (337 nm) was used to generate high-quality MS and MS/ MS data. 9. The oMALDITM Server software (Applied Biosystems, Foster City, CA, USA), a Windows-based program consisting of a main window, a set of drop-down menus, and a series of dialog boxes, controls the oMALDITM ion source to operate the motor controller and the UV laser and directs the acquisition software (Analyst QS, Applied Biosystems) to acquire mass spectra. The oMALDITM Server software with Imaging option allows the user to select a rectangular region of the user-defined area anywhere on the MALDI plate, to build a search pattern covering the defined region according to the spot width and height parameters, to acquire MS or MS/MS spectra from each pixel, and to rapidly obtain two-dimension mode (2D) profile information for small molecules from intact tissue sections.

3. Methods 3.1. Drug Administration and Tissue Sampling

1. Treat two Spray-Dawley rats with astemizole orally at 100 mg per kg of body weight in 0.4% methylcellulose as the dosing vehicle (see Note 1). 2. Attach perfusion needle (20 gauge) with a clamp closed. 3. Fill perfusion reservoir with 50 ml ice cold D-PBS. 4. Anesthetize the rat using CO2 and open the abdominal cavity exposing the heart.

150

Hsieh, Li, and Korfmacher

5. Insert the needle into the left ventricle of the heart and open the clamp to allow the flow of D-PBS. Make a cut in the right atrium with sharp scissors to allow for the escape of blood from the animal. 6. Stop the perfusion when blood has been cleared from body (liver has turned pale in color). 7. Excise the brain from the skull, snap freeze it, and allow the brain immediately to freeze in dry ice to prevent drug diffusion and to keep the integrity of the tissue. 8. Store the frozen tissue in sealed container at –20◦ C until further analysis to prevent the water evaporation from the tissue. 3.2. Tissue Preparation

1. Use ice slush made from distilled water to attach the rat brains to the cryostat sample stages (see Notes 2 and 3). 2. Adjust the chamber temperature of cryostat to –20◦ C. 3. Clean a new slicing blade with methanol. 4. Install the blade on Cryostat. 5. Set the section thickness to 16 µm. 6. Equilibrate steel MALDI sample plate to the chamber temperature of cryostat. 7. Slice the rat brain tissue samples along coronal direction to observe cortex and hypothalamic areas, main binding sites for histamine H-1 receptors, and three brain ventricles, choroids plexus, dorsal third ventricle, and lateral ventricle. These ventricles are important for us to evaluate drug penetration of the brain–blood barrier. Other brain substructures, such as hippocampus, cingulum, and external capsule, were also exposed. 8. Keep slicing the brain tissue until the desired position is reached. 9. Transfer the tissue slice on a cold MALDI plate with an artist brush and move it to the desired location (see Note 4). 10. Put one finger against the opposite side of the MALDI plate to thaw the tissue slice on the MALDI sample plate. 11. Dehydrate the rat tissue sections in a vacuum desiccator at the room temperature overnight before matrix application.

3.3. Matrix Application

1. Place ∼10 ml of a matrix solution into a glass reagent sprayer. 2. Hold the sample plate vertically about 20–30 cm from the sprayer nozzle.

Mapping Pharmaceuticals in Rat Brain Sections

151

3. Spray multiple coats of matrix across the surface of the tissue. One coating cycle consisted of passing the sprayer two times across the surface of the tissue. 4. Allow the sample to dry for at least 2–5 min before the next coating cycle to avoid deposition of a large quantity of solvent in any one region of the tissue. The combination of the spray rate and spray distance should be adjusted to avoid excessive wetting of the tissue which could lead to analyte dispersion (see Notes 5, 6, and 7). 5. Iterates this process to redissolve and recrystallize the matrix enhancing the incorporation efficiency of analytes in the crystals until there was a relatively homogeneous layer of matrix crystals over the surface of the tissue, usually about 15 cycles. 6. Dehydrate the processed tissue sections under the hood for 15 min at room temperature. 3.4. MALDI-MS Method Development

1. Deposit a few drops of matrix solution containing 10 ng/µl astemizole on the MALDI plates. 2. Allow all spots to dry completely under the hood prior to acquiring MS and MS/MS information of astemizole with the oMALDITM Server software (Fig. 8.1) (see Notes 8–10). 3. Load the MALDI plate into the MALDI chamber of a QStar Pulsar hybrid QqTOF mass spectrometer. 4. Operate repetition rates and the pulse energy of the laser at 20 Hz and 80 µJ. m/z 135

Astemizole

1.5e6

135

m/z 308

459

218

OCH3

N

N

1.0e6

N H

N

m/z 218

5.0e5 Intensity, cps

308 100

200

300

F

400

500

600

445

m/z 121 m/z 308

3.0e5

OH

M-14 metabolite N

N

2.0e5

204 N

1.0e5

121 100

N H

m/z 204

308 331 200

300

400

500

600

F

m/z

Fig. 8.1. Product ion spectra of astemizole and its desmethyl metabolite. (Adapted from Li et al. (19) with permission.)

152

Hsieh, Li, and Korfmacher

5. Set argon used as the collision gas with a collision-induced dissociation (CAD) gas pressure setting of 5 (∼3–4 × 10−5 torr). 6. Adjust collision energy for dissociation of the protonated astemizole molecular ion [M+H]+ (parent ion m/z 459 shown in Fig. 8.1) (see Note 11). 7. Acquire m/z 218 product ion scans for astemizole and m/z 204 product ion scans for its demethyl metabolite (Fig. 8.1) under the “enhanced mode” as suggested by high performance liquid chromatography/quadrupole linear ion trap mass spectrometric experiments (Scheme 8.1) to build a MALDI-IMS method for further imaging mass spectrometry studies. The predominant fragments for both astemizole and its primary metabolite were then “enhanced,” using a feature of the Pulsar that decelerates and traps ions prior to accelerating them into the orthogonal time-of-flight region. Sensitivity could be further increased by allowing a wide mass range into the collision quadrupole (∼5 m/z), by using the maximum pulsing rate into the TOF region, and by detecting only a small mass range surrounding the fragment of interest. Profile metabolites in the tissue homogenate using a liquid chromatography /hybrid triple quadrupole linear ion trap mass spectrometry

Acquire mass spectral and structure information of metabolites

Set up MALDI-IMS methods for the primary metabolites

Prepare tissue sections for MALDI-IMS experiments

Map the dosed compound and its metabolites using MALDI-IMS

Scheme 8.1 General strategy of mapping pharmaceuticals and their metabolites.

8. Deposit a few 0.2 µl drops of the matrix solution containing the analyte within the various regions of the blank brain section. 9. Allow all spots to dry under the hood for 1 h. 10. Load the MALDI plate into the MALDI chamber of a QStar Pulsar hybrid QqTOF mass spectrometer.

Mapping Pharmaceuticals in Rat Brain Sections

153

11. Measure the MALDI signals of the analyte from the spotted regions under the established instrumental conditions to evaluate the degree of the ionization suppression due to the endogenous materials from the blank tissues (see Note 12). 12. Record the optical image of a sagittal section of the brain (Fig. 8.2a) to indicate the presence of the cortex, limbic system, cerebellum, brain stem, and ventricles regions and to define the outline of a given tissue section. 3.5. Imaging Astemizole and Its Metabolite in the Rat Brain Sections

1. Load the MALDI plate with the study brain slices coated with DHB into the MALDI chamber of a QStar Pulsar hybrid QqTOF mass spectrometer. 2. Initiate the image process with the oMALDITM Server software. The sample MALDI plate will be moved from one spot to the other spot under a stationary laser automatically, creating a raster of desorbed areas over the tissue surface.

A

B

hip cx

v1 v2 v3

cx

cc

yp

hyp

C

D

Fig. 8.2. (a) The optical image of a rat brain from a coronal section. (b) MALDI-MS/MS images of astemizole in the rat brain slice without perfusion and (c) with perfusion; cortex (cx), hippocampus (hip), corpus callosum (cc), hypothalamic region (hyp), thalamus region (yp), choroid plexus (v1), dorsal third ventricle (v2), and lateral ventricle (v3) are indicated by arrows. (d) MALDI-MS/MS images of M-14 metabolite of astemizole in the rat brain slice. (Adapted from Li et al. (19) with permission.)

154

Hsieh, Li, and Korfmacher

The oMALDI server directs the acquisition software to acquire mass spectra from each point, treating each laser spot as one sample in a batch run. The laser was turned off for 2 s while the MALDI plate was repositioned for the next acquisition location in a rectangular pattern with a spatial resolution of 100–150 µm (the size of the laser beam). During an imaging mass spectrometric experiment, the resulting mass spectra data are first obtained as a function of the acquisition times which are associated with the location in an array of pixels (see Notes 13, 14, 15, and 16). 3.6. Data Analysis

1. Construct the ion density image maps after data acquisition has completed using the oMALDI Server 4.0 imaging software. Two-dimensional images are obtained by plotting the spatial dimensions of x and y versus the signal amplitude of selected ion range as a function of the location on the tissue surface. 2. Construct two-dimensional ion maps for astemizole based on a given product ion m/z 218 value that was monitored in each mass spectrum. As shown in Fig. 8.2, the color of each pixel represents the intensity of the selected ion contributed from astemizole. In order to detect small molecules in a complex biological tissue section using MS/MS was normally required in order to generate signals from the compound of interest that could easily be distinguished from the background interference produced by the matrix. Analysis of the resulting product ions by the second analyzer generates a mass spectrum of the product ions that can be used to provide structural information. Therefore, by monitoring the transition of a selected precursor ion to its product ions, tandem mass spectrometry systems offer a tool for distinguishing between isobaric compounds such as matrix ions and the analytes (see Notes 17 and 18). 3. Reconstruct two-dimensional ion maps for the demethyl metabolite based on a given product ion m/z 204 value that was monitored in each mass spectrum. As shown in Fig. 8.2d, the color of each pixel represents the intensity of the selected ion contributed from the metabolite (see Note 19).

4. Notes 1. The animal dosing experiments were carried out in accordance to the US National Institutes of Health Guide to the Care and Use of Laboratory Animals and the Animal Welfare Act.

Mapping Pharmaceuticals in Rat Brain Sections

155

2. Tissue sample preparation is one of the critical steps for the success of detecting small molecules in biological tissues using MALDI-IMS. Inappropriately handling tissue samples in the sample preparation steps may cause delocalization and degradation of the analytes. Tissue containing drugs subject to photodegradation should be maintained in a dark container as much as possible during the processing. 3. Cryostats are normally used as a standard tool for slicing frozen tissue to reducing sample contamination. Contamination of the tissue surface due to the use of OCT (optimal cutting temperature polymer) as an embedding medium for stabilizing the organ specimens should be avoided because it would lead to ionization suppression in MALDI-MS analysis. 4. There are several other ways available to transfer tissue slices to the sample plate. First, the frozen section could adhere to the MALDI plate held at room temperature by placing the plate over the section. Second, the section is stuck to a double-sided transparent tape which is then stuck on the sample plate. 5. Once the section is mounted to the sample plate with the desired orientation, matrix solution could be also deposited on the tissue surface by electrospray deposition or using robotics to deposit small matrix droplets across the tissue surface before MALDI analysis. 6. High-resolution images with MALDI-IMS necessitate uniform coating by the matrix, no redistribution of surface analytes and linearity of MALDI signals after matrix application. In order to produce a homogeneous crystal layer over the tissue surface, spraying multiple coats of matrix across the surface of the tissue is recommended. 7. For certain analytes, contaminants such as salts and phospholipids in tissues may significantly inhibit co-crystallation of analytes with MALDI matrix resulting in low MALDI intensities of the analytes. Washing of tissue sections with weak organic solvents such as 70% ethanol prior to matrix application might eliminate endogenous materials from the native tissue and further to improve the MALDI matrix crystallization process. However, for small molecule applications, there is a potential risk of either losing the analytes or altering the spatial integrity of localization of drugrelated components on tissues. 8. MALDI-MS employs a matrix solution to mix with the analyte to allow drying and to form co-crystals form. These crystals are subject to absorb energy at the wavelength of

156

Hsieh, Li, and Korfmacher

the laser beam resulting in desorption and ionization of the analytes that were included in the crystals. In this process, matrix is uniformly deposited over various tissue sections to extract analytes into the surface of the tissue and to produce crystals. 9. The selection of matrix and matrix solution conditions such as solvent compositions, pH, and the rates of co-crystals growth can affect the quality of mass spectra for small molecules. The success of the MALDI-IMS applications to an analyte in tissue is strongly dependent on the choice of appropriate matrix materials. 10. The common UV-absorbing molecules used as matrices for MALDI analysis are benzoic acid-based components with low molecular weights (< 500 Da) which dominate the low-mass range background for a typical MALDI-MS spectrum further challenging the advancement of MALDI for the analysis of small molecules. DHB was chosen in this work over sinapinic acid (SA, 3,5-dimethoxy, 4-hydroxy cinnamic acid) commonly used as the matrix because it generated fewer background signals and greater analyte signals for astemizole. 11. The instrument parameters for the analytes are first optimized prior to MALDI-IMS experiments. 12. In general, it is possible for MALDI to suffer some degrees of signal irreproducibility due to crystallization behavior and laser properties such as energy profile and firing repetition rate. Variations in peak intensity of analytes may be seen when the laser focuses on different regions of the same tissue section. To investigate the potential of the ionization reproducibility of astemizole from different regions of rat brain tissue sections, seven drops (0.1 µl each) containing 10 ng of both analytes were deposited on different locations of a blank rat brain slice. The MALDI responses of astemizole from the spiked areas of rat brain sections were found to be comparable. The results confirmed MALDIMS to be a feasible technique in providing spatial resolution with a high degree of consistency to reveal the density and distribution of astemizole and its primary metabolite in rat brain tissue. 13. A raster of organ sections from a rodent species containing the compounds of interest under a stationary laser beam is performed over a predetermined two-dimensional array to generate ion plumes directly from the tissue sections in a MALDI plate array. 14. The movement of the sample stages is automatically accomplished in the x and y directions to locate the edges of

Mapping Pharmaceuticals in Rat Brain Sections

157

the tissue sample and to define the exact region of interest. The localization of the laser beam on the sample is accurate to within 5 µm. 15. MALDI imaging resolution is governed by both crystal size and laser diameter. Generally smaller crystal sizes yield better imaging spatial resolution. With crystal diameters smaller than the laser beam, typically 50–200 µm depending on the instrument, imaging resolution is generally limited to the laser diameter. 16. Acquisition times for tissue imaging relies on several instrumental parameters such as spatial resolution requirements, the laser repetition rate, spot-to-spot sample repositioning transfer time, and data processing. As laser with fast repetition rates and improved electronics become available, one could reduce the acquisition times from hours to minutes. 17. Figure 8.2b shows the MALDI-MS/MS image of astemizole in a rat brain slice. The product ion spectrum of astemizole from the rat brain tissue was found to be consistent with that from the authentic standard material deposited into the blank rat brain section. The MALDI-MS/MS imaging results as given in Fig. 8.2b revealed that the astemizole was readily detected in the entire brain and the most intense signal was observed in the ventricle area. A significant amount of astemizole appeared to be distributed and localized at the cortex, hippocampus, hypothalamic, and thalamus regions over the entire thickness of the section. Astemizole was found to be low in the corpus callosum which contains nerve fibers. The product ion spectrum of astemizole clearly indicated the presence of the drug with variable concentrations as a function of the location within the tissue section monitored. 18. The MALDI-MS/MS imaging results of astemizole shown in Fig. 8.2c indicated that the relative distribution patterns between the perfused and untreated rat brain sections remained consistent over the entire investigation period. The fact that the images were similar despite the sample preparation differences suggested that the drug level in the slices was likely due to tissue binding activities rather than being due to residual biological fluids from the systemic circulation. 19. The MALDI-MS/MS imaging results as given in Fig. 8.2d revealed that the M-14 metabolite was also measured by MS primarily around the three ventricle sites. The product ion spectrum of M-14 metabolite from the rat brain tissue was found to be consistent with that from brain homogenates by HPLC-MS/MS (data not shown). The

158

Hsieh, Li, and Korfmacher

signals of the M-14 metabolite cannot be confirmed in other parts of the brain section. The results implied that the metabolite tended to stay in cerebrospinal fluid than to cross the brain–blood barrier (BBB). Therefore, astemizole was likely the major cause of efficacy and CNS side effects following drug administration in rats. References 1. Simons, F. E. (1994) The therapeutic index of newer H1-receptor antagonists. Clin Exp Allergy, 24, 707–723. 2. Simons, F. E. (1999) Prospective, long-term safety evaluation of the H1-receptor antagonist cetirizine in very young children with atopic dermatitis. ETAC Study Group. Early treatment of the atopic child. J Allergy Clin Immunol, 104, 433–440. 3. Passalacqua, G., Bousquet, J., Bachert, C., Church, M. K., Bindsley-Jensen, C., Nagy, L., Szemere, P., Davies, R. J., Durham, S. R., Horak, F., Kontou-Fili, K., Malling, H. J., Cauwenberge, P., Canonica, G. W. (1996) The clinical safety of H1-receptor antagonists. Allergy, 51, 666–675. 4. Estelle, F., Simons, R. (1999) H1-receptor antagonists: safety issues. Ann Allergy Asthma Immunol, 83, 481–488. 5. Horak, F., Stubner, U. P. (1999) Comparative tolerability of second generation antihistamines. Drug Saf, 20, 385–401. 6. Jansen, F. P., Wu, T. S., Voss, H. P., Steinbusch, H. W., Vollinga, R. C., Rademaker, B., Bast, A., Timmerman, H. (1994) Characterization of the binding of the first selective radiolabelled histamine H3-receptor antagonist, [125I]-iodophenpropit, to rat brain. Br J Pharmacol, 113, 355–362. 7. Alves-Rodrigues, A., Timmerman, H., Willems, E., Lemstra, S., Zuiderveld, O. P., Leurs, R. (1998) Pharmacological characterisation of the histamine H3 receptor in the rat hippocampus. Brain Res, 788, 179–186. 8. Khatib-Shahidi, S., Andersson, M., Herman, J. L., Gillespie, T. A. Caprioli, R. M. (2006) Direct molecular analysis of whole-body animal tissue sections by imaging MALDI mass spectrometry. Anal Chem, 78, 6448–6456. 9. Hsieh, Y., Casale, R., Fukuda, E., Chen, J., Knemeyer, I., Wingate, J., Morrison, R., Korfmacher, W. A. (2006) Matrix-assisted laser desorption/ionization imaging mass spectrometry for direct measurement of clozapine in rat brain tissue. Rapid Commun Mass Spectrom, 20, 965–972. 10. Hsieh, Y., Chen, J., Korfmacher, W. A. (2007) Mapping pharmaceuticals in tissues

11.

12.

13.

14.

15.

16.

17.

18.

19.

using MALDI imaging mass spectrometry. J Pharmacol Toxicol Methods, 55, 193–200. Reyzer, M. L., Hsieh, Y., Ng, K., Korfmacher, W. A., Caprioli, R. M. (2003) Direct analysis of drug candidates in tissue by matrix-assisted laser desorption/ionization mass spectrometry. J Mass Spectrom, 38, 1081–1092. Stoeckli, M., Staab, D., Schweitzer, A., Gardiner, J., Seebach, D. (2007) Imaging of a beta-peptide distribution in whole-body mice sections by MALDI mass spectrometry. J Am Soc Mass Spectrom, 18, 1921–1924. Chaurand, P., Norris, J. L., Cornett, D. S., Mobley, J. A., Caprioli, R. M. (2006) New developments in profiling and imaging of proteins from tissue sections by MALDI mass spectrometry. J Proteome Res, 5, 2889–2900. Chaurand, P., Sanders, M. E., Jensen, R. A., Caprioli, R. M. (2004) Proteomics in diagnostic pathology: profiling and imaging proteins directly in tissue sections. Am J Pathol, 165, 1057–1068. Chaurand, P., Schwartz, S. A., Billheimer, D., Xu, B. J. Crecelius, A. Caprioli, R. M. (2004) Integrating histology and imaging mass spectrometry. Anal Chem, 76, 1145–1155. Chaurand, P., Schwartz, S. A., Caprioli, R. M. (2002) Imaging mass spectrometry: a new tool to investigate the spatial organization of peptides and proteins in mammalian tissue sections. Curr Opin Chem Biol, 6, 676–681. Chaurand, P., Schwartz, S. A., Caprioli, R. M. (2004) Assessing protein patterns in disease using imaging mass spectrometry. J Proteome Res, 3, 245–252. Chen, J., Hsieh, Y., Crossman, L., Knemeyer, I., Korfmacher, W. A. (2008) Visualization of first-pass drug metabolism of terfenadine by MALDI-imaging mass spectrometry. Drug Metab Lett, 2, 1–4. Li, F., Hsieh, Y., Kang, L., Sondey, C., Lachowicz, J., Korfmacher, W. A. (2009) MALDI-tandem mass spectrometry imaging of astemizole and its primary metabolite in rat brain sections. Bioanalysis, 1, 299–307.

Chapter 9 Laser Ablation Electrospray Ionization for Atmospheric Pressure Molecular Imaging Mass Spectrometry Peter Nemes and Akos Vertes Abstract Laser ablation electrospray ionization (LAESI) is a novel method for the direct imaging of biological tissues by mass spectrometry. By performing ionization in the ambient environment, this technique enables in vivo studies with potential for single-cell analysis. A unique aspect of LAESI mass spectrometric imaging (MSI) is depth profiling that, in combination with lateral imaging, permits 3D molecular imaging for the first time under native conditions. With current lateral and depth resolutions of ∼100 and ∼40 µm, respectively, LAESI MSI helps to explore the molecular architecture of live tissues. Key words: Mass spectrometry, imaging, ambient, direct analysis, depth profiling, threedimensional, in vivo, tissue imaging.

1. Introduction Traditional mass spectrometric imaging (MSI) methods, such as matrix-assisted laser desorption ionization (MALDI) and secondary ion mass spectrometry (SIMS), have become important tools for the investigation of molecular distributions in tissues due to their high ionization efficiencies and excellent lateral and depth resolutions. Invasive sample preparation and the need for vacuum conditions, however, are incompatible with the analysis of live samples. Novel ionization methods in ambient mass spectrometry (1) overcome these limitations by performing imaging under native conditions. Desorption electrospray ionization (2), atmospheric pressure (AP) mid-infrared (mid-IR) MALDI (3), laser ablation S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_9, © Springer Science+Business Media, LLC 2010

159

160

Nemes and Vertes

electrospray ionization (LAESI) (4), and most recently laser ablation coupled to flowing AP afterglow MS (5) have demonstrated imaging capabilities with lateral resolutions 20–400 µm while obtaining low limits of detection. Figure 9.1a shows the schematics of the LAESI ion source. Figure 9.1b depicts the fast imaging of the entrainment of laser-ablated particulates into the electrospray plume. The interaction of the sprayed droplets with the particulates and neutrals emerging from the laser ablation (LA) produces coalesced charged particles that are thought to be the basis of the LAESI signal. In this chapter, we describe the protocols for lateral and 3D MSI of biological tissues using LAESI.

(A)

(B)

Fig. 9.1. (a) Schematics of the LAESI setup with (optional) spray current measurement and fast imaging system (C, capillary emitter; SP, syringe pump; HV, high-voltage power supply; L-N2 , nitrogen laser; M, mirrors; FL, focusing lenses; CV, cuvette; CE, counter electrode; OSC, digital oscilloscope; SH, sample holder; L-Er:YAG, Er:YAG laser; MS, mass spectrometer; PC-1 to PC-3, personal computers). Open black dots represent the droplets formed by the electrospray. Their interaction with the particulates and neutrals (solid gray dots) emerging from the laser ablation produces some fused particles (solid black dots) that are thought to be the basis of the LAESI signal. (Adapted with permission from (6). Copyright 2007 American Chemical Society.) (b) Our fast imaging experiments supported this scenario. The image captured with ∼10 ns exposure time shows the interaction of the laser ablation plume (LA) and the electrospray plume (ES). (Adapted with permission from (6). Copyright 2007 American Chemical Society.)

2. Materials 2.1. Reagents and Sample

1. Electrospray solution for positive ion mode analysis: 50% methanol or acetonitrile, 1% acetic acid or formic acid in water. CAUTION: Glacial acetic acid is extremely harmful when inhaled; causes burns to the skin and eyes; and avoid contact with skin, eyes, and the respiratory system. 2. Electrospray solution for negative ion mode analysis: 50% methanol or acetonitrile, 1% ammonium acetate or ammonium hydroxide in water. CAUTION: May cause irritation to skin, eyes, and the respiratory tract; avoid direct contact and inhalation.

Laser Ablation Electrospray Ionization

161

3. Electrospray solution for reactive LAESI experiments: reagents, e.g., ∼1 µM–1 mM lithium sulfate. 4. Pre-cleaned glass slides (e.g., Thermo Fisher Scientific, Inc., Waltham, MA, USA). 5. Double-sided tape. 6. 1-Methyl-2-pyrrolidinone and an electric heater for the removal of the optical fiber jacket and 1% nitric acid (reagent grade) for the etching of the germanium oxide fiber. CAUTION: Combustible, causes irritation to skin and eyes; and avoid contact with skin and eyes. 7. For plant studies, growth chamber and plant growing protocol (e.g., 14 h photoperiod, 24◦ C). 8. Cryomicrotome with microtome blades (e.g., Shandon Cryostat, Thermo Fisher Scientific), frozen specimen embedding medium (FSEM, Shandon Cryomatrix, Thermo Fisher Scientific, Inc., Waltham, MA, USA), and liquid N2 . CAUTION: It causes suffocation when present at amounts sufficient to reduce oxygen concentration below 19.5%. Contact with tissue can cause severe cryogenic burns. Always handle with protective gloves. 9. –80◦ C freezer (e.g., Revco freezers, Thermo Fisher Scientific, Inc., Waltham, MA, USA). 10. Aluminum foil. 2.2. Mid-IR Ablation on Sample

1. Q-switched mid-IR laser emitting light at 2.94 µm wavelength or an optical parametric oscillator tunable in the vicinity of 2.94 µm wavelength pumped by a Q-switched Nd:YAG laser (e.g., Opotek, Inc., Carlsbad, CA, USA). CAUTION: Class IV laser; Direct exposure of the eye to the laser beam can cause permanent eye damage. Always wear laser protective eyewear of sufficient optical density at the operating wavelengths. 2. Mirrors for mid-IR light (e.g., gold-coated mirrors, Thorlabs, Newton, NJ, USA). 3. Focusing lens for mid-IR light, e.g., plano-convex CaF2 or antireflection-coated ZnSe lens (Infrared Optical Products, Farmingdale, NY, USA); or an aspherical lens; or a reflective microscope objective (Newport Corp., Irvine, CA, USA); or a sharpened optical fiber for mid-IR light delivery, e.g., 450 µm core germanium oxide-based optical fiber (Infrared Fiber Systems, Silver Spring, MD, USA) with chucks and positioners (e.g., Newport Corp., Irvine, CA, USA) and a three-axis translation stage (e.g., Thorlabs, Inc., Newton, NJ, USA).

162

Nemes and Vertes

4. Thermo- and photosensitive papers for beam alignment and core spot size optimization (e.g., R30C5W, Liquid Crystal Resources L.L.C., Glenview, IL, and multigrade IV, Ilford Imaging Ltd, UK), respectively. 5. Plate holders (e.g., FP02 or FP02, Thorlabs, Inc., Newton, NJ, USA) for sample mounting. 6. Enclosure for the ion source including the laser ablation plume. CAUTION: Ablated particles may become airborne and pose health hazard upon inhalation or contact with skin and eye. Please always be aware of related health risks and take proper measures for protection. 7. For frozen samples Peltier cooling stage (e.g., Ferrotec Corp., Bedford, NH, USA) with heat fan (e.g., Allied Electronics, Inc., Fort Worth, TX, USA), and DC power supply. 8. Optical microscope for measurement of ablation crater dimensions. 2.3. Post-ionization with Charged Droplets

1. Electrospray emitters (e.g., MT320-50-5-5 or FS360-7530-N-5, New Objective, Inc., Woburn, MA, USA). 2. Metal union, conductive perfluoroelastomer ferrule, fittings, tubing sleeve, fused silica capillary, needle port (e.g., IDEX Health & Sciences, Oak Harbor, WA, USA, or Waters Corp., Milford, MA, USA). 3. Syringe pump (e.g., Harvard Apparatus, Holliston, MA, USA) or an LC solvent-delivery system. 4. For experiments with controlled spaying mode, stainless steel electrode with oscilloscope to perform spray current measurements (e.g., WaveSurfer 452, LeCroy, Chestnut Ridge, NY, USA). 5. High-voltage power supply (e.g., Stanford Research Systems, Inc., Sunnyvale, CA, USA). CAUTION: Electric shock hazard. Please make sure that all electric connections are properly shielded.

2.4. Molecular Imaging and Data Analysis

1. Three-axis translation stage (e.g., Newport Corp., Irvine, CA, USA) with motorized actuators and controller (e.g., with LTA-HS, Newport Corp., Irvine, CA, USA). 2. Mass spectrometer (e.g., Q-TOF Premier, Waters Corp., Milford, MA, USA). 3. Software for correlated positioning of the three-axis translation stage with laser ablation and mass spectrometric data acquisition (written in, e.g., LabView, National Instruments, Austin, TX, USA).

Laser Ablation Electrospray Ionization

163

4. Software for data analysis and scientific visualization package for molecular image generation (e.g., Origin, OriginLab Corp., Northampton, MA, USA); ImageJ (NIH, available at http://rsb.info.nih.gov), Biomap (available at http://www.maldi-msi.org).

3. Methods Figure 9.2a shows the workflow of LAESI imaging in an MSI setup. In a 2D imaging experiment, the sample surface is scanned in the focal plane of the mid-infrared laser light, the ablated neutrals and particulate matter are post-ionized with a cloud of charged droplets, and the resulting ions are mass analyzed and recorded. Mass-selected molecular images are reconstructed by correlating the intensity of ion signal for an m/z of interest with the absolute lateral coordinate of analysis for each pixel of the interrogated area (see Fig. 9.2b for an Aphelandra squarrosa leaf and Fig. 9.4b,c for rat brain tissue section). For 3D molecular imaging, the chemical depth profile of the sample is acquired with a selected number of laser pulses delivered at each pixel. The 3D molecular image is represented by a set of 2D images correlated with the absolute depth of analysis (see Fig. 9.2c). The imaging resolution of LAESI MSI laterally is characterized by the diameter of the ablation crater along with the ablation depth in the third dimension. In regular experiments, lateral step sizes, larger than or equivalent to the diameter of the ablation crater, are applied. The molecular imaging resolution is therefore limited by the divergence of the laser and the properties of the optical focusing component. High-resolution LAESI MSI is also dependent on reliable and effective ion generation at each pixel of the image. Governing factors are the efficiency of the interaction between the electrospray and the laser ablation plumes and the success of entrainment of the resulting droplets and ions into the mass spectrometer. Figure 9.3a summarizes the major variables that require optimization for a robust operation. The following section is guidance for setting up a LAESI MSI experiment. For best results, please consider the average ion count for at least three pixels each time adjustment is made to a variable. 3.1. Preparation and Mounting of Tissue

1. Mount sample directly on sample holder with the surface of interest exposed on top and adjust dOR-FP to ∼10 mm (see Fig. 9.3a). Small clamps and single- and double-sided adhesion tape often enable easy mounting. CAUTION:

164

Nemes and Vertes

Fig. 9.2. (a) Schematics of LAESI MSI in two and three dimensions. Molecular images are obtained by correlating the coordinates of the analyzed pixels with the selected ion abundances. (b) Examples for metabolites with uniform (m/z 663.16 assigned as kaempferol-(diacetyl coumaryl-rhamnoside)) and heterogeneous (m/z 493.10 assigned as methoxy kaempferol glucoronide) lateral distributions in A. squarrosa leaves. (Reprinted with permission from (4). Copyright 2008 American Chemical Society.) (c) 3D imaging revealed that kaempferol or luteolin detected at m/z 287.06 (yellow scale) followed the yellow variegation sectors whereas chlorophyll a with m/z 893.54 (blue scale) accumulated in the third and fourth layers of the leaf. (Reprinted with permission from (11). Copyright 2009 American Chemical Society.)

Native water content of sample must be retained for midIR light coupling (see Note 4.1.1). (Optional) Evaporative loss of water can be prevented by thaw mounting, whereby the sample is directly frozen onto the sample holder or a glass slide and is kept at low temperatures during imaging (see Fig. 9.4a).

Laser Ablation Electrospray Ionization

165

Fig. 9.3. (a) Geometry parameters in a LAESI MSI experiment used to optimize signalto-noise ratio and lateral and depth resolution of molecular images. The relative position of the electrospray capillary, ES, and the focal point, FP, with respect to the mass spectrometer orifice, OR, and the incidence angle, α, of the laser beam are the essential factors that determine the efficiency of interaction between the ablation plume and the electrospray. (b) The smaller droplets generated by a low flow rate nanospray source (compare the ion intensities at 25 and 300 nl/min) and careful alignment of the ablation plume improved the sensitivity of LAESI MSI. (c) Differential interference contrast microscope image of the adaxial surface of Arabidopsis thaliana leaf ablated using a plano-convex ZnSe focusing lens showed tissue removal in an area of ∼200 µm diameter suggesting a similar lateral resolution for the MSI experiment. (d) Online visualization of tissue sampling with a fiber sharpened to ∼50 µm (see the top left corner of the image) further improved the lateral resolution by yielding ablation marks of ∼100 µm in diameter.

2. (Optional) Sectioning may be required for certain tissue types such as animal or human organs. Wrap sample in aluminum foil and dip it into liquid N2 for ∼20 s. CAUTION: Improper timing can lead to tissue fracturing. Section tissue into 10–100 µm thick slices with a cryomicrotome at –10 to –20◦ C. Note that temperature and time requirements might depend on the tissue type. Thaw-mount slices onto microscope slides or directly onto sample holder. 3. (Optional) Store tissues and sections wrapped in aluminum foil at –80◦ C. This maintains sample integrity for up to a few months. CAUTION: Improper storage conditions can result in postmortem tissue degradation (see Note 4.1.2). 3.2. Mid-IR Ablation of Sample

1. For lateral imaging experiment, operate a mid-IR laser at 2.94 µm wavelength and 10 Hz repetition rate. (Optional)

166

Nemes and Vertes

Fig. 9.4. LAESI MSI experiment on ∼100 µm thick coronal section of a rat (Rattus norvegicus) brain. The sample was mounted on a Peltier cooling stage and was kept frozen during imaging to avoid postmortem tissue degradation. (a) The imaged area is shown by an array of ablation marks located 200 µm apart in the x and y directions. Scale bar corresponds to 1 mm. (b) Glycerophosphocholine observed at m/z 258.11 exhibited heterogeneous distribution in the tissue. (c) Glycerophosphocholine (38:6) measured at m/z 806.56 was present throughout the section and appeared especially abundant in the caudate putamen (striatum) and cerebral cortex regions of the brain section.

For samples with low water content (e.g., skin) increase the incident pulse energy. For very low water content samples (such as bone and tooth), tune the laser wavelength to a strong absorption band (see Note 4.2.1). 2. Use gold mirrors and a focusing element (e.g., a planoconvex ZnSe lens) to couple the mid-IR laser pulse into the sample at right angle (α=90◦ in Fig. 9.3a) (see Note 4.2.2). Position the focal point below the orifice axis and set dOR-FP to 5–8 mm (see Fig. 9.3a) (see Note 4.3.3). CAUTION: Protective enclosure must be used to avoid inhalation of airborne ablated particles. 3. Optimize the position of the focusing element and the pulse energy of the laser beam to achieve tissue removal in desired dimensions – this determines the pixel size (see Note 4.2.2). For example, Fig. 9.3c shows that a ZnSe lens allowed the sampling of Arabidopsis thaliana with 200 µm diameter ablation mark at 1 mJ/pulse energy. (Optional) Fine focusing can also be achieved by vertical positioning of the

Laser Ablation Electrospray Ionization

167

sample holder (see SH in Fig. 9.3a). Please remain within 5 < dOR-FP < 30 mm range (see Note 4.3.3). At distances less than 5 mm the expansion of the ablation plume can destabilize the electrospray (6). Most of the ablated particulate matter does not travel beyond 30 mm due to the drag force in the atmosphere (7). 4. (Optional) Sharpened optical fiber delivery can offer superior lateral resolutions by producing ablation craters with decreased dimensions. For example, Fig. 9.3d shows a 100-µm diameter ablation crater. For a GeO2 -based fiber delivery system, first place ∼1 cm of both tips of the fiber in boiling 1-methyl-2-pyrrolidinone for ∼5 min until the polyimide coating is dissolved. CAUTION: Work under hood with good ventilation to avoid inhalation of vapor. Wash residues off the glass fiber tip with water and methanol. Cleave both ends for flat surfaces with a diamond scribe and etch one end to desired diameter in 0.1% HNO3 solution at room temperature (e.g., 50 µm in ∼15 min). Clean fiber tip with water. Position fiber tip to surface using a threeaxis translation stage. CAUTION: Etched tip poses sharp object hazard; handle with care. 3.3. Post-ionization with Charged Droplets

1. Position a blunt-tip nanospray emitter (e.g., TaperTipTM with 100 µm inner diameter) in line with the inlet axis of the mass spectrometer and at orifice-to-emitter tip distance, dOR-ES , of ∼10 mm (see Fig. 9.3a). 2. Feed 50% methanol solution containing 0.1% acetic acid or 0.1% ammonium acetate for positive and negative ion mode, respectively, through the metal emitter at 10 µm). Metabolite imaging within cellular organelles, which requires resolution at submicrometer level, will provide researchers with valuable information about cellular metabolism. In this regard, imaging with SIMS has achieved submicron spatial resolution and has successfully visualized subcellular structures (6, 7) and dynamic changes of metabolite molecules during Tetrahymena mating (6).

Matrix-Assisted Laser Desorption/Ionization

175

Table 10.2 Current IMS for lipids

Glycerophospholipids

Neutral lipids

Sphingolipids

Fatty acids

Class

Ion detection mode

PCs

Positive

References (1, 12, 27, 34, 43)

PEs

Negative

(27, 43, 44)

PIs

Negative

(27, 43, 44)

PSs

Negative

(27, 43, 44)

PGs

Negative

(27, 43, 44)

Cardiolipins Triacylglycerols

Negative Positive

(27, 43, 44) (43)

Diacylglycerols

Positive

(43)

Cholesterol

Positive

(27, 46)

Gangliosides

Negative

(12, 28)

Sulfatides

Negative

(28, 32, 44)

Galactosylceramide

Positive

(9, 47)

–

Negative

(48)

Note: PC=phosphatidylcholine, PE=phosphatidylethanolamine, PI=phosphatidylinositol, PS=phosphatidylserine, PG=phosphatidylglycerol.

However, both soft ionization of analytes and tandem MS are difficult to achieve with typical SIMS technique (8). In contrast, spatial resolution of MALDI-IMS is lower than that of SIMS. The spatial resolution depends on the experimental conditions and the instrument used but is typically 20–100 µm. Limitations of the spatial resolution of MALDI-IMS include the size of the organic matrix crystal and the analyte migration during the matrix application process. To overcome these problems, Taira and colleagues reported a nanoparticle (NP)-assisted laser desorption/ionization (nano-PALDI)-based IMS, in which the matrix crystallization process is eliminated (9). The use of nano-PALDI has enabled researchers to image compounds with spatial resolution at the cellular level (15 µm; almost equal to the size of the diameter of a laser spot). Here, practical experimental procedures using MALDI-IMS for lipid metabolites are presented. The technical problems specific to IMS of small metabolites are introduced and the techniques for mitigating these problems are explained. In addition, as an attractive alternative to MALDI-IMS, nanoparticle-based IMS, which improves spatial resolution, is discussed. A protocol for the preparation of extremely small nanoparticles (d = 3.7 nm) developed by our group (10) and its application method is presented. Before describing the detailed experimental procedure, the structure of glycerophospholipids (GPLs) should be briefly

176

Sugiura and Setou

mentioned. GPLs are the most abundant type of lipids, especially in the brain, and form a large molecular family in which phosphoric acid in the ester form is bound to a glycerolipid. They are subdivided into distinct classes (e.g., phosphatidylcholines [PCs], phosphatidylethanolamines, and phosphatidylinositols) based on the structure of the head group linked to the phosphate, which is attached at the sn-3 position of the glycerol backbone. GPLs are further subdivided into numerous molecular species based on the composition of the fatty acids linked to the sn-1 and sn-2 positions of the glycerol backbone (Fig. 10.1).

Fig. 10.1. Structure of phospholipids. Structure of the glycerol backbone of glycerophospholipids (GPLs) (a) and their head group (b) is shown. GPLs are subdivided into distinct classes (e.g., phosphatidylcholines, phosphatidylethanolamines, and phosphatidylinositols) based on the structure of the head group linked to the phosphate, which is attached at the sn-3 position of the glycerol backbone (b). GPLs are further subdivided into numerous molecular species based on the composition of fatty acids linked to the sn-1 and sn-2 positions of the glycerol backbone.

Matrix-Assisted Laser Desorption/Ionization

177

IMS allows the visualization of these multiple molecular species in parallel. The distinct localization of GPL molecular species can be determined; in other words, the distinct fatty acid composition of biological membranes in different tissue locations can be determined.

2. Materials 2.1. Chemicals

1. Trifluoroacetic acid (TFA). 2. 1,2-Dihexanoyl-sn-glycero-3-phosphocholine (MW: 453.5, monomer and dimer were used for calibration) as a calibration standard. 3. 2,5-Dihydroxybenzoic acid (DHB).

2.2. Matrix Solution for MALDI-IMS of PCs

1. For measurement of PC, a DHB solution (40 mg/ml DHB, 10 mM potassium acetate, 70% methanol) (see Notes 1 and 2) was used as the matrix solution (11).

2.3. Matrix Solution for MALDI-IMS of Gangliosides

1. For measurement of gangliosides, a matrix solution without potassium salt (40 mg/ml DHB, 70% methanol) was used as the matrix solution (12).

2.4. Chemicals for Nanoparticle Preparation

1. FeCl2 ·4H2 O (99.9%). 2. γ-Aminopropyltriethoxysilane (γ-APTES). All of the chemicals used in this study were of the highest purity available.

3. Methods 3.1. Animal Sacrifice and Tissue Extraction

All experiments involving mice were conducted in accordance with the protocols approved by the animal care and use committee at the participating research institute. 1. The brains of 8-week-old male C57BL/6 J Cr mice were used. 2. The brains were extracted within 1 min (typically 40 s) after sacrifice (see Note 3). 3. The trimmed tissue blocks were immediately frozen in powdered dry ice, which allows tissues to be frozen without cracks, and stored at –80◦ C until use.

178

Sugiura and Setou

3.2. Preparation of Tissue Sections

1. Tissues blocks were sectioned at –16◦ C using a cryostat (CM 3050; Leica, Germany) to a thickness of 5 µm (for detailed discussion about slice thickness, see Refs. (13, 14)). 2. Tissue blocks were held by an optimum cutting temperature (OCT) polymer, but they were not embedded into the polymer because it was thought that any residual polymer on the tissue slices might degrade the mass spectra (14) (see Note 4 and Ref. (14)). 3. The frozen sections were thaw-mounted on indium-tinoxide (ITO)-coated glass slides (Bruker Daltonics) and ITOcoated sheets (Tobi Co., Ltd., Kyoto, Japan). The slides were used for tandem time-of-flight (TOF/TOF) measurements and the sheets were used for quadrupole ion trap (QIT)-TOF measurements. 4. The prepared sections were subjected to matrix application within 5 min (see Note 5).

3.3. Spray Coating of the Matrix Solution

1. The matrix solution was sprayed over the tissue surface using a 0.2-mm nozzle caliber airbrush (Procon Boy FWA Platinum; Mr. Hobby, Tokyo, Japan). 2. The distance between the nozzle tip and the tissue surface was held at 10 cm and the spraying period was fixed at 5 min. Approximately 100 µl of matrix solution was sprayed onto each brain section.

3.4. Production of Nanoparticles

200 ml is the best quantity for the preparation of nanoparticles. Day 1 1. Cool water and ethanol at 4ºC temperature. 2. Dissolve 2 g of FeCl2 ·4H2 O in 100 ml water. 3. Decant 100 ml γ-APTES directly into another beaker. γ-APTES dissolves in both water and organic solvents. 4. Agitate γ-APTES using a stir bar. 5. Pour all the solution of FeCl2 into the γ-APTES solution with agitation. The mixing ratio of γ-APTES to FeCl2 ·4H2 O should be 45:1. 6. Put an aluminum foil lid on the beaker and allow it to cool for 30 min. 7. Separate the solution into 35 ml portions in 50-ml Falcon tubes and centrifuge the solution at 8,000×g force at 4ºC for 20 min. 8. Remove the supernatant. 9. Add chilled water so that the total volume is 35 ml.

Matrix-Assisted Laser Desorption/Ionization

179

10. Break the precipitate with supersonic waves and vortex it to suspend the precipitate. 11. Centrifuge the solution at 8,000×g force at 4ºC for 20 min. 12. Remove the supernatant. 13. Repeat procedures 9 through 12, three times. 14. Add ethanol, centrifuge the solution, and remove the supernatant (Fig. 10.2a). 15. Leave the tube overnight at 80ºC to completely evaporate the ethanol.

Fig. 10.2. Precipitated nanoparticles by centrifugation (a) and after drying/crushing in a mortar (b).

Day 2 1. Crush the precipitate using a mortar as shown in Fig. 10.2b. 2. After crushing, collect the powder from the crushed precipitate and put it in Eppendorf tubes. 3. The prepared nanoparticles can be preserved for 1 month in a dark room at 4ºC. 3.5. Use of Nanoparticles for Imaging Mass Spectrometry

1. Put 10 mg each of nanoparticles into two Eppendorf tubes. 2. Put 1 ml of 100% methanol with 10 mM sodium acetate into each tube. 3. Vortex the contents for 1 min with supersonic waves. 4. Centrifuge the tubes at 9,000×g at 4ºC for 10–30 s. 5. Adjust the centrifuging time so that the ideal color of the fluid can be obtained (Fig. 10.3). 6. Transfer the supernatant, which contains the nanoparticles, to fresh Eppendorf tubes. 7. Spray the supernatant onto the tissue sections using an airbrush. Adjust the distance between the airbrush and the tissue samples so that the sprayed droplets of methanol evaporate during spraying or so that they dry immediately on the tissue surface.

180

Sugiura and Setou

Fig. 10.3. A suspension of nanoparticles.

3.6. Instrument Parameter Settings

IMS was performed using a MALDI TOF/TOF instrument (Ultraflex 2 TOF/TOF; Bruker Daltonics), which was equipped with a 355 nm Nd:YAG laser. 1. The number of laser irradiations was 100 in each spot. 2. Raster scans on tissue surfaces were performed automatically using FlexControl and FlexImaging 2.0 software (Bruker Daltonics). 3. Image reconstruction was performed using FlexImaging 2.0 software. 4. For measurement of PCs, the data were acquired in the positive reflectron mode under an accelerating potential of 20 kV using an external calibration method. 5. For measurement of gangliosides, experiments were performed in the negative reflectron mode under an accelerating potential of –20 kV.

3.7. Spectrum Normalization

The variation in ionization efficiency in the IMS results, caused by heterogeneous distribution of organic matrix crystals and their sublimation during the measurement, was eliminated for each data point by equalizing the total ion current for each mass spectrum, using the “normalize spectra” function of the FlexImaging 2.0 software (11) (see Note 6).

3.8. Tandem Mass Spectrometry for Molecular Identification

Molecular identification was performed with tandem mass spectrometry using a QIT-TOF mass spectrometer (AXIMA-QIT; Shimadzu, Kyoto, Japan) to ensure the molecular assignment which was performed using only mass. The MSn analysis was performed directly on the sections of tissue sections in the mid-mass range mode.

Matrix-Assisted Laser Desorption/Ionization

3.9. Results of Imaging Mass Spectrometry for Phosphatidylcholines

181

As a number of studies have shown, MALDI-IMS is effective for profiling and visualizing the distribution of GPLs because of easy detection. GPLs are ionized efficiently because of the following reasons. First, large amounts (more than 60% dry weight) of mouse brain are composed of lipids (high expression). Second, GPLs have an easily ionized structure. Phospholipids, in particular PCs, contain a phosphate group and a trimethylamine that are charged easily (15). Figure 10.4a shows representative mass spectrum obtained by IMS of mouse brain sections. Nine intense mass peaks were assigned to eight abundant PC molecular species and one sphingomyelin based on m/z values. All of these contain a trimethylamine head group. In Fig. 10.4b, the tissue distributions of the five major PC molecular species in sections of the sagittal brain are shown. Although the most abundant molecular PC species, PC (diacyl-16:0/18:1), was uniformly distributed across the entire gray matter region of the section, other PC molecular species showed rather heterogeneous distribution patterns.

Fig. 10.4. Differential distribution of phosphatidylcholine (PC) molecular species in sagittal mouse brain sections. (a) An averaged mass spectrum obtained from an entire mouse brain section. In the spectrum, intense mass peaks corresponding to nine abundant PCs were assigned according to mass. (b). MALDI-IMS of a brain section simultaneously identified the heterogeneous distributions of several PCs. Schema of the mouse brain sagittal section and ion images of PCs obtained by IMS are shown.

182

Sugiura and Setou

Fig. 10.5. Tandem mass spectrometry (MS) allows molecular identification of glycerophospholipids on the tissue surface. Results of MSn structural analysis of ions corresponding to the PCs. Both MS2 and MS3 product ion spectra show that the mass peaks are derived from the PCs. Neutral losses of 59 and 124 u from precursor ions, corresponding to trimethylamine and cyclophosphate, respectively, were used as diagnostic ions.

Matrix-Assisted Laser Desorption/Ionization

183

The molecular assignments were verified by structural analysis of each peak using MSn (Fig. 10.5). A QIT-TOF mass spectrometer was used for this purpose. This instrument can identify molecules using a highly selective/sensitive MSn from mixture ions generated on the tissue surface (5). In each mass peak, the presence of a trimethylamine head group and a phosphate was confirmed (neutral losses of 59 and 124 u from precursor ions corresponding to trimethylamine and cyclophosphate, respectively), which are used as diagnostic ions in product ion mass spectra (5, 16, 17). 3.10. Results of Imaging Mass Spectrometry for Gangliosides

Gangliosides are glycosphingolipids consisting of mono- to polysialylated oligosaccharide chains of variable lengths attached to a ceramide unit. Gangliosides are inserted in the outer layer of plasma membranes with the hydrophobic ceramide moiety acting as an anchor. Their oligosaccharide moiety is exposed to the external medium (18). Gangliosides also form a large family; their constituent oligosaccharides differ in the glycosidic linkage position, sugar configuration, and the content of neutral sugars and sialic acid. In particular, based on the number of sialic acids, they can be subdivided into GM (i.e., mono-sialylated), GD (di-sialylated), GT (tri-sialylated), and GQ (quadra-sialylated) groups. As well as the oligosaccharide unit, the ceramide moiety also varies with respect to the type of long-chain base (LCB; sphingosine base) and fatty acid moiety (Fig. 10.6).

Fig. 10.6. Structures of ganglioside molecular species containing C18-long-chain backbone (LCB) and C20-LCB. The C20 species has two more carbons in its LCB moiety than does the C18 species (arrow).

Previous biochemical studies have shown that the LCB of brain gangliosides has either 18 or 20 carbons (i.e., C18- or C20-sphingosine). The C20-sphingosine (C20-LCB species) is present in significant amounts only in the central nervous system (19–22). Its content increases significantly throughout life in rodents and in human (23–25). Because of its characteristic brain specificity and the dramatic increase during the course of life, the

184

Sugiura and Setou

C20-LCB ganglioside was of great interest. However, a lack of visualization technology for the specific detection and visualization of C18 and C20 gangliosides has prevented researchers from determining their precise tissue distribution. Antibodies against some oligosaccharide moieties are available for visualizing molecular species with different constituent oligosaccharides (26); however, such immunological methods cannot detect differences in the ceramide structure, which is hidden in the lipid bilayer. Because of the negatively charged sialic acids and their rich abundance in the brain, gangliosides are strongly detected in the 1,500 < m/z < 2,500 range in the negative ion detection mode (12, 27, 28) (Table 10.3). In addition, as well as structural differences in oligosaccharides, IMS can discriminate structural differences in lipid moieties and has successfully revealed the specific distribution of C20-LCB species in mouse brain. Although C18 species are widely distributed throughout the frontal brain, C20 species are selectively localized in specific brain regions, namely in the molecular layer of the dentate gyrus (Fig. 10.7A).

Table 10.3 Detection of gangliosides in the mouse hippocampus Negative ions [M–H]–

[M+Na– 2H]–

[M+K– 2H]–

[M+Na– 3H]–

[M+Na+K3H]–

[M+2 K– 3H]–

GM1 (d18:1/18:0)

1,544

–

–

–

–

–

GM1 (d20:1/18:0)

1,572

–

–

–

–

–

GD1 (d18:1/18:0)

–

1,858

1,874

–

–

–

GD1 (d20:1/18:0)

–

1,886

1,902

–

–

–

GT1 (d18:1/18:0)

–

–

–

2,170

2,186

2,202

GT1 (d20:1/18:0)

–

–

–

2,198

2,214

2,230

◮ Fig. 10.7. (continued) spectra of m/z 888.3 and 916.3 were obtained to determine the different structural constituents in the ceramide moieties. Because of the detection of m/z 283.0 (fatty acid-related ion) in both spectra, the 28 u difference between m/z 1,544 and m/z 1,572 was attributed to the difference in the sphingosine constituent (m/z 1,544 had C18 sphingosine and m/z 1,572 had C20 sphingosine).

Matrix-Assisted Laser Desorption/Ionization

185

Fig. 10.7. Imaging mass spectrometry (IMS) and direct MSn allow specific visualization and detection of ganglioside molecular species. (A) IMS at 50 µm raster step size was used to gain an overview of ganglioside distribution in different brain regions. Schematic diagram of the brain section (a) and ion images of sulfatides (b–c) are shown. Ions corresponding to gangliosides, namely GD1 (d–i) and GM1 (j–l), were visualized. (B) (a) The MS2 product ion spectra show that the ions at m/z 1,544 and 1,572 had the same oligosaccharide structure (i.e., they contained a sialic acid moiety) but the ceramide mass peaks were observed at different m/z values. (b) MS3 product ion mass

186

Sugiura and Setou

To confirm that the difference of 28 u, which corresponds to a (CH2 )2 unit, observed between C18 and C20 species (Table 10.3) can be attributed to differences in LCB chain lengths, structural analysis of ions corresponding to GM1 gangliosides was performed using MSn (Fig. 10.7B). This technique can provide detailed structural information about the ion of interest. The MS2 results for both m/z 1,544 and 1,572 showed a ceramide peak and peaks corresponding to oligosaccharides containing a sialic acid (Fig. 10.7B). This 10.7B, a). The peaks in the MS2 spectra for oligosaccharides of m/z 1,544 and 1,572 were exactly the same. These gangliosides have the same oligosaccharide moiety. Next, MS3 was performed to determine the detailed structure of the ceramide. In the MS3 spectra, a common peak was observed at m/z 283.0, which corresponded to (C17 H35 COOH)– , a fatty acid (Fig. 10.7B, b). Thus the difference was because of the difference in the chain length of the LCBs, namely the C18 and C20 sphingosines. 3.11. Results of NanoparticleAssisted Laser Desorption/Ionization Imaging Mass Spectrometry

In the early study of the soft ionization, NPs (d = ∼30 nm) with metal oxide cores were found to assist laser desorption/ionization of analyte molecules in the presence of diluted glycerol (29). Recently, gold-NPs (d = ∼5.5 nm) have been used in MS (30) and IMS (31). Compared to the traditional MALDI-IMS, nanoparticle-based IMS has several advantages. Gold-NPs ionize biomolecules that are different from those by traditional organic matrices (31) (in the results shown here, NPs more easily ionize sphingolipids than GPLs). Also, elimination of matrix-derived signals is important, especially for analysis of small molecules (9, 32). Elimination of the matrix–analyte co-crystallization process is another important advantage of NP-based IMS. Spatial resolution is not restricted by the crystal size but by the diameter of the laser spot. Figure 10.8 shows the results of NP and MALDI-IMS of lipids in rat cerebellum. As can be seen, spraying NPs on the tissue surface did not alter the optical image of the tissue structure (Fig. 10.8, upper panel), and it enhanced soft ionization of lipids to visualize them (Fig. 10.8, middle panel). In contrast, when the sections were sprayed with a DHB solution, although shown is a poor example, non-homogeneous crystals were observed on the section, which obscured the optical view of the sample surface (Fig. 10.8, upper panel). This resulted in blurred images of ions in this section (Fig. 10.8, middle panel). Furthermore, soft ionization with NP achieves sufficient ion yields to perform MS/MS on the tissue surface (Fig. 10.8, bottom). Because of the extremely small particle size, it is possible to localize lipids in fine tissue structures (within several tens of micrometers). Figure 10.9 shows another example of nanoparticle-IMS used to visualize the distribution of lipids

Matrix-Assisted Laser Desorption/Ionization

187

Fig. 10.8. Nanoparticle-assisted laser desorption/ionization imaging of lipids. (Upper panel) Optical images of rat cerebellum tissue before/after being sprayed with nanoparticles (NPs) and 2,5-dihydroxybenzoic acid (DHB) solution were shown. Successive brain section stained with hematoxylin–eosin (H&E) is also presented. (Middle panel) Ion images obtained with NPs and DHB are shown. Visualized ions were identified as galactosylceramide (C24h:0) and PC (diacyl-34:2) by tandem mass spectrometry on both DHB- and NP-coated sections. (Bottom panel) Example representation of MS/MS result on the tissue sprayed with NP, for ion at m/z 850.8. Product ion spectrum indicates that the ion was derived from galactosylceramide(C24h:0).

188

Sugiura and Setou

Fig. 10.9. Nanoparticle-assisted laser desorption/ionization improves spatial resolution in imaging mass spectrometry. (Left panel) Optical image of rat hippocampus indicating measurement area for nanoparticle-based IMS. Nissl-stained section indicating fine layer structure of rat hippocampus is also shown. (Right panel) Ion images which reveal hippocampal layer-specific distribution of phosphatidylinositol (18:0/20:4) at m/z 885.5 and sulfatide (24:1) at m/z 888.8 were presented. GCL = granular cell layer, IML = inner molecular layers, MML = middle molecular layers. Reprinted from Ref. (9).

within the layer structure of the rat hippocampus. In this case, hippocampal layer-specific localization of sulfatides and phosphatidylinositol was clearly observed (33).

4. Notes 1. The composition of organic solvents (methanol) used in matrix solutions influences the signal detection sensitivity for lipids and peptides (Fig. 10.10). Higher composition of methanol enhances lipid detection by efficient extraction of lipid molecules from tissue sections. In contrast, hydrophilic

Matrix-Assisted Laser Desorption/Ionization

189

Fig. 10.10. Composition of organic solvents (methanol) in the matrix solution influences signal detection of lipids and peptides. Matrix solutions (containing 35 mg/ml DHB and 0.1% TFA) of different ratios of water and methanol were prepared. Next, 0.5 µl of each solution was applied to the section of mouse brain homogenate, which has homogeneous molecular distribution at any location of the section (n = 3). By increasing the methanol concentration, crystal form was also changed; needle-like crystal, from which peptides were detected, was changed to the aggregate of smaller crystals from which lipids were detected.

solutions enhance peptide detection for the same reason. Considering the easier application of methanol by spraying, a solution of 70% methanol was used in this study. 2. GPLs, especially PCs, preferentially cationize as their alkali metal-adduct molecules (34–36). Because tissue sections contain rich sodium and potassium salts, such alkali metal-adducted GPLs, rather than protonated molecules, are preferentially generated. Molecular ionization with such multiple ion forms from a single species often hampers IMS experiments, because GPLs have many molecular species and a single peak might contain multiple types of ions. Generation of such multiple molecular ions from a single PC molecular species can be suppressed by adding potassium acetate to the matrix solution (1). In this study, potassium salt (20 mM potassium acetate) was added to the matrix solution. Thus, the molecular ion forms were limited to potassium-adducted molecules, and the spectra were simplified. 3. As shown in Fig. 10.11 postmortem degradation of GPLs was observed by IMS within 15 min in a series of mouse brains extracted at different times (15, 30, 60, and 120 min).

190

Sugiura and Setou

Fig. 10.11. Postmortem degradation of phosphatidylcholines and increase in lysophosphatidylcholines. An IMS series was performed on mouse brains extracted after different amounts of time had elapsed after sacrifice (within 1, 15, 30, 60, and 120 min after sacrifice). After IMS, the ion intensities of the PCs were averaged over the entire section. As postmortem events, degradation of PCs and an increase in lyso-PCs were observed within 15 min, presumably because of the stimulation of phospholipase A under ischemic conditions (37, 38). In this study, mouse brains were extracted within 1 min (typically within 40 s) after sacrifice. Reprinted from Ref. (11).

This is presumably because of stimulation of phospholipase enzymes under ischemic conditions (37, 38). 4. In general, embedded tissues are cut into thin slices. Embedding enables the sample to retain its shape and makes the cutting process easier. However, in IMS experiments, attachment and penetration of the embedding agents (e.g., OTCs) in the sample lead to a deterioration of the MS signal (4, 14). In particular, when analyzing small molecules with an m/z of 800–2,000, contamination with OTC leads to the presence of extremely high polymer peaks in the mass spectra of positive ions. This virtually hides all of the smaller peaks (Fig. 10.12). For this reason, when preparing sections for IMS, OTC is used only to “support” the tissue blocks, and thus, OTC does not directly attach to the tissue being analyzed. As an alternative, a precooled semiliquid gel of 2% sodium carboxymethyl cellulose (CMC) can be used as an embedding compound that does not interfere with mass spectrometry (39). 5. Dehydration of tissue sections for long times can lead to altered signals (40). Goodwin and colleagues demonstrated that, even within 1 min, signals were altered, both increasing and decreasing. Therefore, tissue slices should be moved to the next step (matrix application) as quickly as possible. Considerable care is required at these stages in order to facilitate a comparison between the biomarkers in independent IMS experiments.

Matrix-Assisted Laser Desorption/Ionization

191

Fig. 10.12. Mass spectra of small molecules were obtained from a representative section that was completely embedded in optimal cutting temperature compound. Reprinted from Ref. (11).

6. The matrix–analyte crystallization process, particularly when using salt-added matrix solution to reduce the molecular ion forms of GPLs, leads to the development of heterogeneous crystals which in turn results in spot-to-spot variance of signal intensities (41, 42). This problem was solved using a spectrum normalization procedure with TIC (Fig. 10.13). We performed spectrum normalization with TIC; the obtained spectra were multiplied with arbitrary variables such that all spectra had equal TIC values (i.e., equal integral values of the measured m/z region [400 < m/z < 900]). Such TIC normalization is available with the “normalize spectra” function of FlexImaging 2.0 software with filter function to exclude a number of noise spectra from the normalization process (see details in the software manual). To evaluate the effect of the normalization procedure, we prepared a section of mouse brain homogenate that had a uniform distribution of biomolecules. Figure 10.13a shows the ion images for m/z 772.6 corresponding to PC (diacyl16:0/16:0), with and without spectrum normalization. After the normalization procedure, the image was corrected such that the ion distribution was uniform throughout the section. The signal intensity was then plotted and found to have a Gaussian distribution. Spectrum normalization with TIC improved the results of the IMS of mouse brain sections. Figure 10.13b shows the ion images of a mouse brain section for PC (diacyl-16:0/16:0),

192

Sugiura and Setou

Fig. 10.13. Spectrum normalization using TIC improves both the quantitative ability and the visualization quality of IMS. (a) IMS results for PC (diacyl-16:0/16:0) on a section of mouse brain homogenate, processed with/without TIC normalization (upper panel), and plot of ion-intensity distribution for PC (diacyl-16:0/16:0) obtained from a brain homogenate section, with/without TIC normalization (lower panel). (b) Ion images of PC (diacyl-16:0/16:0) on an adult mouse brain section, in which spectra were processed with/without TIC normalization. Reprinted from Ref. (11).

with and without spectrum normalization. In the ion image without normalization, the ion distribution was heterogeneous, even between adjacent pixels. Furthermore, the signal intensity was found to decrease with time (arrowhead). In contrast, when the normalization procedure was used, a clear ion-distribution pattern that correlated well with the anatomical features of the brain section was obtained (11).

Acknowledgments The authors would like to thank Dr. S. Taira and Dr. H. Ageta for their advice and fruitful discussion. This work was supported by the SENTAN program of the Japan Science and Technology Agency.

Matrix-Assisted Laser Desorption/Ionization

193

References 1. Garrett, T. J., Prieto-Conaway, M. C., Kovtoun, V., Bui, H., Izgarian, N., Stafford, G. Yost, R. A. (2006). Imaging of small molecules in tissue sections with a new intermediate-pressure MALDI linear ion trap mass spectrometer. Int J Mass Spectrom, 260, 11. 2. Khatib-Shahidi, S., Andersson, M., Herman, J. L., Gillespie, T. A., Caprioli, R. M. (2006). Direct molecular analysis of whole-body animal tissue sections by imaging MALDI mass spectrometry. Anal Chem, 78, 6448–6456. 3. Stoeckli, M., Staab, D., Staufenbiel, M., Wiederhold, K. H., Signor, L. (2002). Molecular imaging of amyloid beta peptides in mouse brain sections using mass spectrometry. Anal Biochem, 311, 33–39. 4. Chaurand, P., Norris, J. L., Cornett, D. S., Mobley, J. A., Caprioli, R. M. (2006). New developments in profiling and imaging of proteins from tissue sections by MALDI mass spectrometry. J Proteome Res, 5, 2889–2900. 5. Shimma, S., Sugiura, Y., Hayasaka, T., Zaima, N., Matsumoto, M., Setou, M. (2008). Mass imaging and identification of biomolecules with MALDI-QIT-TOF-based system. Anal Chem, 80, 875–885. 6. Ostrowski, S. G., Van Bell, C. T., Winograd, N., Ewing, A. G. (2004). Mass spectrometric imaging of highly curved membranes during Tetrahymena mating. Science, 305, 71–73. 7. Monroe, E. B., Jurchen, J. C., Lee, J., Rubakhin, S. S., Sweedler, J. V. (2005). Vitamin E imaging and localization in the neuronal membrane. J Am Chem Soc, 127, 12152–12153. 8. Kraft, M. L., Weber, P. K., Longo, M. L., Hutcheon, I. D., Boxer, S. G. (2006). Phase separation of lipid membranes analyzed with high-resolution secondary ion mass spectrometry. Science, 313, 1948–1951. 9. Taira, S., Sugiura, Y., Moritake, S., Shimma, S., Ichiyanagi, Y., Setou, M. (2008). Nanoparticle-assisted laser desorption/ionization based mass imaging with cellular resolution. Anal Chem, 80, 4761–4766. 10. Moritake, S., Taira, S., Sugiura, Y., Setou, M., Ichiyanagi, Y. (2009). Magnetic nanoparticle-based mass spectrometry for the detection of biomolecules in cultured cells. J Nanosci Nanotechnol, 9, 169–176. 11. Sugiura, Y., Konishi, Y., Zaima, N., Kajihara, S., Nakanishi, H., Taguchi, R., Setou, M. (2009). Visualization of the cell-selective distribution of PUFA-containing phosphatidyl-

12.

13.

14.

15. 16.

17.

18.

19.

20. 21.

22.

cholines in mouse brain by imaging mass spectrometry. J Lipid Res, 50, 1776–1788 Hayasaka, T., Goto-Inoue, N., Sugiura, Y., Zaima, N., Nakanishi, H., Ohishi, K., Nakanishi, S., Naito, T., Taguchi, R., Setou, M. (2008). Matrix-assisted laser desorption/ionization quadrupole ion trap timeof-flight (MALDI-QIT-TOF)-based imaging mass spectrometry reveals a layered distribution of phospholipid molecular species in the mouse retina. Rapid Commun Mass Spectrom, 22, 3415–3426. Sugiura, Y., Shimma, S., Setou, M. (2006). Thin sectioning improves the peak intensity and signal-to-noise ratio in direct tissue mass spectrometry. J. Mass Spectrom Soc Jpn, 54, 4. Schwartz, S. A., Reyzer, M. L., Caprioli, R. M. (2003). Direct tissue analysis using matrix-assisted laser desorption/ionization mass spectrometry: practical aspects of sample preparation. J Mass Spectrom, 38, 699–708. Pulfer, M., Murphy, R. C. (2003). Electrospray mass spectrometry of phospholipids. Mass Spectrom Rev, 22, 332–364. Stubiger, G., Pittenauer, E., Allmaier, G. (2008). MALDI seamless postsource decay fragment ion analysis of sodiated and lithiated phospholipids. Anal Chem, 80, 1664–1678. Touboul, D., Piednoel, H., Voisin, V., De La Porte, S., Brunelle, A., Halgand, F., Laprevote, O. (2004). Changes of phospholipid composition within the dystrophic muscle by matrix-assisted laser desorption/ionization mass spectrometry and mass spectrometry imaging. Eur J Mass Spectrom (Chichester, Eng), 10, 657–664. van Echten, G., Sandhoff, K. (1993). Ganglioside metabolism. enzymology, topology, and regulation. J Biol Chem, 268, 5341–5344. Sonnino, S., Chigorno, V. (2000). Ganglioside molecular species containing C18- and C20-sphingosine in mammalian nervous tissues and neuronal cell cultures. Biochim Biophys Acta, 1469, 63–77. Sambasivarao, K., McCluer, R. H. (1964). Lipid components of gangliosides. J Lipid Res, 15, 103–108. Schwarz, H. P., Kostyk, I., Marmolejo, A., Sarappa, C. (1967). Long-chain bases of brain and spinal cord of rabbits. J Neurochem, 14, 91–97. Jungalwala, F. B., Hayssen, V., Pasquini, J. M., McCluer, R. H. (1979). Separation of molecular species of sphingomyelin

194

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

Sugiura and Setou by reversed-phase high-performance liquid chromatography. J Lipid Res, 20, 579–587. Palestini, P., Sonnino, S., Tettamanti, G. (1991). Lack of the ganglioside molecular species containing the C20-long-chain bases in human, rat, mouse, rabbit, cat, dog, and chicken brains during prenatal life. J Neurochem, 56, 2048–2050. Palestini, P., Masserini, M., Sonnino, S., Giuliani, A., Tettamanti, G. (1990). Changes in the ceramide composition of rat forebrain gangliosides with age. J Neurochem, 54, 230–235. Mansson, J. E., Vanier, M. T., Svennerholm, L. (1978). Changes in the fatty acid and sphingosine composition of the major gangliosides of human brain with age. J Neurochem, 30, 273–275. Kotani, M., Kawashima, I., Ozawa, H., Terashima, T., Tai, T. (1993). Differential distribution of major gangliosides in rat central nervous system detected by specific monoclonal antibodies. Glycobiology, 3, 137–146. Jackson, S. N., Wang, H. Y., Woods, A. S. (2005). Direct profiling of lipid distribution in brain tissue using MALDI-TOFMS. Anal Chem, 77, 4523–4527. Chen, Y., Allegood, J., Liu, Y., Wang, E., Cachon-Gonzalez, B., Cox, T. M., Merrill, A. H., Jr., Sullards, M. C. (2008). Imaging MALDI mass spectrometry using an oscillating capillary nebulizer matrix coating system and its application to analysis of lipids in brain from a mouse model of Tay-Sachs/Sandhoff disease. Anal Chem, 80, 2780–2788. Tanaka, K., Ido, Y., Akita, S., Yoshida, Y., Yoshida, T. (1987). Proceedings of the 2nd Japan–China Joint Symposium on Mass spectrometry , Osaka, Japan, 185–187. McLean, J. A., Stumpo, K. A., Russell, D. H. (2005). Size-selected (2–10 nm) gold nanoparticles for matrix assisted laser desorption ionization of peptides. J Am Chem Soc, 127, 5304–5305. Jackson, S. N., Ugarov, M., Egan, T., Post, J. D., Langlais, D., Albert Schultz, J., Woods, A. S. (2007). MALDI-ion mobility-TOFMS imaging of lipids in rat brain tissue. J Mass Spectrom, 42, 1093–1098. Ageta, H., Asai, S., Sugiura, Y., GotoInoue, N., Zaima, N., Setou, M. (2009). Layer-specific sulfatide localization in rat hippocampus middle molecular layer is revealed by nanoparticle-assisted laser desorption/ionization imaging mass spectrometry. Med Mol Morphol, 42, 16–23. Ageta, H., Asai, S., Sugiura, Y., Goto-Inoue, N., Zaima, N., Setou, M. (2009). Layerspecific sulfatide localization in rat hip-

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

pocampus middle molecular layer is revealed by nanoparticle-assisted laser desorption/ionization imaging mass spectrometry. Med Mol Morphol, 42, 16–23. Jackson, S. N., Wang, H. Y., Woods, A. S. (2005). In situ structural characterization of phosphatidylcholines in brain tissue using MALDI-MS/MS. J Am Soc Mass Spectrom, 16, 2052–2056. Han, X., Gross, R. W. (2001). Quantitative analysis and molecular species fingerprinting of triacylglyceride molecular species directly from lipid extracts of biological samples by electrospray ionization tandem mass spectrometry. Anal Biochem, 295, 88–100. Hsu, F. F., Turk, J. (2001). Structural determination of glycosphingolipids as lithiated adducts by electrospray ionization mass spectrometry using low-energy collisionalactivated dissociation on a triple stage quadrupole instrument. J Am Soc Mass Spectrom, 12, 61–79. Umemura, A., Mabe, H., Nagai, H., Sugino, F. (1992). Action of phospholipases A2 and C on free fatty acid release during complete ischemia in rat neocortex. Effect of phospholipase C inhibitor and N-methylD-aspartate antagonist. J Neurosurg, 76, 648–651. Rehncrona, S., Westerberg, E., Akesson, B., Siesjo, B. K. (1982). Brain cortical fatty acids and phospholipids during and following complete and severe incomplete ischemia. J Neurochem, 38, 84–93. Stoeckli, M., Staab, D., Schweitzer, A. (2006). Compound and metabolite distribution measured by MALDI mass spectrometric imaging in whole-body tissue sections. Int J Mass Spectrom, 260, 195–202. Goodwin, R. J., Dungworth, J. C., Cobb, S. R., Pitt, A. R. (2008). Time-dependent evolution of tissue markers by MALDI-MS imaging. Proteomics, 8, 3801–3808. Sugiura, Y., Shimma, S., Setou, M. (2006). Two-step matrix application technique to improve ionization efficiency for matrixassisted laser desorption/ionization in imaging mass spectrometry. Anal Chem, 78, 8227–8235. Mock, K. K., Sutton, C. W., Cottrell, J. S. (1992). Sample immobilization protocols for matrix-assisted laser-desorption mass spectrometry. Rapid Commun Mass Spectrom, 6, 233–238. Astigarraga, E., Barreda-Gomez, G., Lombardero, L., Fresnedo, O., Castano, F., Giralt, M. T., Ochoa, B., RodriguezPuertas, R., Fernandez, J. A. (2008). Profiling and imaging of lipids on brain and

Matrix-Assisted Laser Desorption/Ionization liver tissue by matrix-assisted laser desorption/ionization mass spectrometry using 2-mercaptobenzothiazole as a matrix. Anal Chem, 80, 9105–9114. 44. Jackson, S. N., Wang, H. Y., Woods, A. S. (2007). In situ structural characterization of glycerophospholipids and sulfatides in brain tissue using MALDI-MS/MS. J Am Soc Mass Spectrom, 18, 17–26. 45. Wang, H. Y., Jackson, S. N., Woods, A. S. (2007). Direct MALDI-MS analysis of cardiolipin from rat organs sections. J Am Soc Mass Spectrom, 18, 567–577. 46. Altelaar, A. F., Klinkert, I., Jalink, K., de Lange, R. P., Adan, R. A., Heeren, R. M., Piersma, S. R. (2006). Gold-enhanced

195

biomolecular surface imaging of cells and tissue by SIMS and MALDI mass spectrometry. Anal Chem, 78, 734–742. 47. Cha, S., Yeung, E. S. (2007). Colloidal graphite-assisted laser desorption/ionization mass spectrometry and MSn of small molecules. 1. Imaging of cerebrosides directly from rat brain tissue. Anal Chem, 79, 2373–2385. 48. Zhang, H., Cha, S., Yeung, E. S. (2007). Colloidal graphite-assisted laser desorption/ionization MS and MS(n) of small molecules. 2. Direct profiling and MS imaging of small metabolites from fruits. Anal Chem, 79, 6575–6584.

Chapter 11 Cellular Imaging Using Matrix-Enhanced and Metal-Assisted SIMS A.F. Maarten Altelaar and Sander R. Piersma Abstract Imaging mass spectrometry (IMS) allows the direct investigation of both the identity and the spatial distribution of the entire molecular content directly in tissue sections, single cells, and many other biological surfaces. We describe here the steps required to retrieve the molecular information from tissue sections using matrix-enhanced (ME) and metal-assisted (MetA) secondary ion mass spectrometry (SIMS). Surface metallization by plasma coating enhances desorption/ionization of membrane components such as lipids and sterols in imaging time-of-flight (ToF) SIMS of tissues and cells. High-resolution images of cholesterol and other membrane components can be obtained for single neuroblastoma cells and reveal subcellular details. Alternatively, in ME-SIMS, 2,5-dihydroxybenzoic acid electrosprayed on neuroblastoma cells allows intact molecular ion imaging of phosphatidylcholine (PC) and sphingomyelin (SM) at the cellular level. Key words: Imaging mass spectrometry, lipids, neuroblastoma cells, gold coating.

1. Introduction In SIMS (1) the sample surface is bombarded with a high-energy primary ion beam between 1 and 25 kiloelectronvolts (keV). Typical primary ions used in SIMS include Ga+ , Cs+ , and In+ , with Ga+ being able to provide the smallest probe size (less than 10 nm). Although SIMS does not routinely yield intact protein and peptide signals directly from tissue sections, it does have several advantages compared to MALDI. The most important advantage of SIMS over MALDI is the chemical imaging capabilities routinely delivering submicron spatial resolution (2). Furthermore, the SIMS technique is very sensitive and remarkably S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_11, © Springer Science+Business Media, LLC 2010

197

198

Altelaar and Piersma

versatile since it can analyze almost any kind of solid surface (3). The bombardment of these solid surfaces with high-energy primary ions will induce damage over a certain depth in the sample, resulting in changes in the molecular structure of the constituents in this area. To prevent imaging of the induced damage, SIMS imaging experiments are conducted in two different regimes. In the dynamic regime the entire sample surface is eroded in time and the complete top monolayer is removed. Dynamic SIMS is primarily used in quantitative elemental imaging (4–6) (not the topic of this chapter). In the static SIMS regime a much lower primary ion dose is used compared to dynamic SIMS resulting in less than 1% of the top surface atoms and molecules to interact with the primary ion beam and no primary ion strikes again the damaged region. Consequently, in static SIMS significantly less fragmentation of the molecular content occurs, which allows the technique to be used in imaging of small organic components (7–9). In order to enhance the ionization yield for large intact molecular ions by SIMS, different kinds of surface modifications (MALDI matrices (10–14), silver (15), and gold (16–20)) as well as the use of polyatomic primary ion beams (21–26) have been suggested. Although these methods have shown to be able to desorb and ionize peptide and proteins from model samples, in direct tissue analysis they are highly biased toward lipids and steroids. One explanation for this phenomenon is the surface sensitivity of the technique. Since with SIMS only the top few monolayers are sampled, the technique favors the ionization of compounds with surface propensity like cholesterol and lipids, which are highly abundant in tissue sections. In this chapter surface modifications in SIMS, such as metalassisted (MetA) and matrix-enhanced (ME) SIMS, are described for the ionization of intact biomolecular ions, increasing the applicability of SIMS to real-world biological problems. In MetASIMS a very thin layer (∼1 nm) of a metal (e.g., gold) is deposited on the sample surface to assist in the desorption/ionization process (3, 17, 18). One crucial factor in this method seems the migration of the analytes on the gold surface. In a recent study we have shown that SIMS signals for both the cholesterol and the lipid phosphatidylcholine (PC) increased when these species were deposited on a thin layer of gold. Increased signals for cholesterol were exclusively obtained when the layer of gold was deposited on top of the cholesterol and PC sample (27). The same effect was observed in direct MetA-SIMS tissue (27) analysis as well as in a MetA-SIMS study of dyes by Adriaensen et al. (16). Using ME-SIMS we have demonstrated the possibility of obtaining peptide signals, from a nervous tissue extract from the pond snail Lymnaea stagnalis, identical to those obtained with MALDI-MS, up to m/z 2,590 (10). The analysis readily identified five known peptides with ME-SIMS using 2,5-dihydroxybenzoic

Cellular Imaging Using Matrix-Enhanced and Metal-Assisted SIMS

199

acid as matrix. In ME-SIMS one is not dependent on the migration of the analyte molecules onto the matrix-covered surface, like in MetA-SIMS. The matrix is deposited in a wet environment, and the analyte molecules are extracted from the tissue surface. Still the biases for species with surface propensity remain since during drying these species are pushed to the outside of the forming matrix crystals. Furthermore, sample preparation (described in the protocol) is more stringent since the size of the matrix crystals determines the obtainable spatial resolution.

2. Materials 2.1. Chemicals

1. α-Cyano-4-hydroxycinnamic acid. 2. 2,5-Dihydroxybenzoic acid. 3. Trifluoracetic acid. 4. HPLC-grade water. 5. Ethanol. 6. Gelatine. 7. Liquid isopentane. 8. Dry ice.

2.2. Instruments and Materials

1. Physical Electronics (Eden Prairie, MN, USA) TRIFTII time-of-flight SIMS (ToF-SIMS) equipped with an 115 In+ liquid metal ion gun. 2. Dissection microscope allowing low (10×/40×) (CETI, Antwerpen, Belgium).

magnification

3. Cryomicrotome; Leica CM 3000 cryostat (Leica Microsystems, Nussloch, Germany). 4. −80◦ C freezer. 5. Desiccator containing a silica gel canister. 6. Optical microscope (Leica DMRX) equipped with a digital camera (Nikon DXM1200). 7. Chromatography sprayer, 10 ml (Sigma-Aldrich; Z529710). 8. 9 mm screw top vials, 12×32 mm (Waters, P/N 186000272). 9. Quorum Technologies (Newhaven, East Sussex, UK) SC7640 sputter coater equipped with a gold target (SC510– 314A), a FT7607 quartz crystal microbalance stage, and a FT7690 film thickness monitor.

200

Altelaar and Piersma

10. Conductive glass slides; 25×50×1.1 mm unpolished float glass, SiO2 passivated/indium-tin-oxide coated, Rs = 6±2  (Delta Technologies; CG-40IN-1115). 11. Male Wistar rats (Crl:WU) weighing 350 g (Charles-River, Maastricht, The Netherlands). 12. Freshwater snails Lymnaea stagnalis, raised under laboratory conditions. 2.3. Software

1. Both TRIFT systems are operated by WinCadence software (version 3.7.1.5) and controlled by the vacuum watcher (Physical Electronics, Watcher 2.1.2.140). 2. AcqirisLive 2.11 controls the acqiris settings and data acquisition. 3. LaVision, DaVis 6.2.3. controls the CCD camera settings and data acquisition. 4. Mass spectral data analysis is performed with WinCadence 3.7.1.5, MatLab 7.0.4 (PCA), AWE3D 1.5.2.0, and tofToCsv (tool to convert entire m/z data file to a comma-separated file (csv)). 5. Image data analysis is performed with WinCadence 3.7.1.5, MatLab 7.0.4 (PCA) and PCA-based in-house-developed algorithms.

3. Method The described protocol for IMS allows the mapping of molecular distributions directly in tissue sections and on cell surfaces. The different approaches to SIMS are able to deliver images of distributions of small organic compounds like lipids and steroids. Since ME-SIMS and especially MetA-SIMS are capable of imaging biological surfaces with very high spatial resolution, these techniques could be very well suited for direct analysis of lipid distributions on single-cell surfaces. 3.1. Tissue Sections

All experimental procedures should be in accordance with the European directives (86/609/EEC) and approved by the Commission on Laboratory Animal Experiments. Male Wistar rats (Crl:WU) weighing 350 g are decapitated without prior anesthesia, and brains are dissected and frozen in liquid isopentane, cooled to −50◦ C on dry ice, and then stored at −80◦ C until sectioning. Freshwater snails (Lymnaea stagnalis) raised under laboratory conditions; 20±1◦ C water temperature, 12 h light/12 h dark

Cellular Imaging Using Matrix-Enhanced and Metal-Assisted SIMS

201

cycle, and fed lettuce ad libitum. Adult specimens are decapitated, and the cerebral ganglia are dissected, directly embedded in 10% gelatine at 30◦ C, and frozen at −80◦ C. Gelatine embedding allows cryostat sectioning down to 5 µm thickness and does not interfere with ME-SIMS analysis. Sections are thaw-mounted on indium-tin-oxide (ITO, 4–8  resistance)-coated glass slides and are stored at −80◦ C until use. Conservation of morphology was checked by optical microscopy during the sectioning, drying, and storage process. No obvious ice crystal damage was observed after freezing at −80◦ C in gelatine. Light microscopy images were acquired using a microscope with a digital camera. Prior to mass spectrometry, frozen tissue sections were brought to room temperature in a desiccator over a silica gel canister. 1. Freeze tissue directly in liquid isopentane after dissection, cool on dry ice, and store at −80◦ C until use. Postmortem changes are known to occur in the proteome of susceptible peptides and proteins within minutes. To prevent these alterations of the sample, the dissected tissue must be snap-frozen and defrosting of the sample should be done only just before sample preparation starts (see Note 1). 2. Place the dissected tissue on the sample holder in the cryostat. Tissue section can be attached to the sample holder by a very small amount of Tissue-Tek. Great care has to be taken that the Tissue-Tek does not come into contact with the tissue area of interest (see Note 2). 3. For very small organs (like the pituitary gland or a mollusk central nervous system) embedding in non-polymer containing solutions like 10% gelatine helps to prevent damaging of the tissue and assist in cutting. Tissue is embedded in 10% gelatine at 30◦ C directly after dissection and frozen at −80◦ C. Gelatine embedding allows sectioning down to 5 µm thickness results in no tissue damage during freezing and is compatible with MS. 4. Cut tissue section ∼10-µm thick using a cryomicrotome at −20◦ C. 10-µm thickness is optimal for IMS; enough analyte molecules available for extraction and no problems with conductivity are observed (see Note 3). 5. Pick up the tissue sections with microscope glass slides by thaw-mounting, place them in a closed container (e.g., plastic Petri dishes sealed with parafilm) on dry ice, and store at −80◦ C until use. 6. Take a tissue section in its closed container from the −80◦ C storage and allow it to come to room temperature in a dry box with silica gel canisters before analysis.

202

Altelaar and Piersma

3.2. Cell Cultures

1. Cells are grown directly on conductive ITO-coated glass slides by seeding (neuroblastoma) cells in six-well plates at ∼25,000 cells/well and cultured in 3 ml of DMEM supplemented with 10% serum and antibiotics. 2. The cells were washed in 300 mM sucrose solution (to prevent hypotonic shock). 3. Next the sucrose is removed by washing with Milli-Q water, after which the cells are rapidly frozen on dry ice. 4. The snap-frozen cells are freeze-dried for 30 min, to remove any ice crystals, and stored at −80◦ C until use. 5. Prior to mass spectrometry, the cell cultures are brought to room temperature in a desiccator over a silica gel canister (1 h). 6. Conservation of cell morphology is checked by optical microscopy, using a Leica DMRX microscope with a Nikon DXM1200 digital camera.

3.3. Matrix Deposition for ME-SIMS

The matrix deposition method of choice depends on the spatial resolution required. For high spatial resolution IMS, electrospray deposition (ESD) is preferred since it results in considerably smaller crystal sizes than pneumatically assisted nebulization (e.g., by a TLC sprayer). Key issues in development of a matrix deposition method are optimal incorporation of analyte into the matrix crystals and minimal lateral diffusion. These two requirements can be met if the matrix arrives at the tissue surface in very small droplets before all solvent has evaporated. 1. For ESD, a syringe pump pumps matrix solution (15 mg/ml 2,5-DHB in 50% MeOH/0.1% TFA (V/V)) from a gastight syringe through a stainless steel electrospray capillary (o.d. 220 µm, i.d. 100 µm) maintained at 3−5 kV (see Note 4). 2. The capillary is mounted on an electrically insulated manual translation stage in a vertical orientation. The stage is fitted with a digital micrometer for accurate positioning of the needle tip with respect to the grounded sample plate. 3. The sample plate is mounted on an X−Y moveable table. 4. Matrix deposition was performed by spraying for 10 min at a flow rate of 12 µl/h, a voltage of 4.7 kV, and a needle to sample plate distance of 5.0 mm. 5. For ME-SIMS analysis, samples are covered with a thin layer of matrix by electrospray deposition yielding small (0.3−1 µm) matrix crystals (see Note 5). 6. After matrix deposition the matrix coverage is checked using an optical microscope and the tissue sections are left to dry for 30 min.

Cellular Imaging Using Matrix-Enhanced and Metal-Assisted SIMS

3.4. Gold Coating for MetA-SIMS

203

1. Place the ITO slide with the sample in the Quorum Technologies SC7640 sputter coater. Press the start sequence button to pump down, flush with argon, further pump down the vacuum chamber, and leak in argon until the pressure reaches 0.1 mbar (all done automatically). 2. Enter on the film thickness monitor the density of the metal used (19.30 g/cm3 for gold) and the desired thickness of the sputtered metal layer. 3. Coat the tissue sections with 1 nm of gold directly on the tissue. 4. Put the discharge voltage on 1 kV and press start on the film thickness monitor. Adjust the plasma current to 25 mA for homogenous coverage. IMS experiments can be performed using ME-SIMS or MetA-SIMS. In the cases described here, these experiments are conducted on a ToF-SIMS equipped with an 115 In+ liquid metal ion gun. The experimental procedure for ME-SIMS and MetASIMS data acquisition is identical.

3.5. ME-SIMS and MetA-SIMS Imaging Mass Spectrometry

All static SIMS experiments were performed on a time-of-flight SIMS (ToF-SIMS) equipped with an 115 In+ liquid metal ion gun. In short, the TRIFT is a stigmatic imaging ToF analyzer incorporating a 2-m flight path and three quasi-hemispherical electrostatic sector analyzers (ESAs) as an integral part of the ToF analyzer. These ESAs compensate for the differences in ion flight times due to variations in the secondary ion’s initial kinetic energies and ion emission angles. The secondary ions were extracted through a 3.2-kV electric field into the ToF analyzer and postaccelerated by an additional 5 kV prior to detection on a dual multichannel plate/phosphor screen detector. A multistop time-todigital converter with 138-ps time resolution was used to acquire the detector signals. All experiments were performed with a primary ion beam current of 450 pA, a primary pulse length of 30 ns, a spot diameter of 500 nm, and a primary ion energy of 15 kV. The experimental conditions were chosen in such a way that all the analyses were conducted in the static SIMS regime. For each chemical image, the primary beam was rastered over a 150 × 150 µm sample area, divided into 256 × 256 square pixels. The positive ion mode mass spectra and chemical images were taken from the same 150 × 150 µm areas within the sample. For each ion detected, the mass (ToF) and primary ion beam position were recorded, allowing post-processing of the data. The chemical images were compared to optical images of the corresponding sample areas. 1. Before conducting SIMS imaging experiments optimize the setup for image quality, using a copper grid with a 25 µm

204

Altelaar and Piersma

repeat. In the vacuum watcher close spectro gate valve (V5). In WinCadance software go to hardware, start the DC beam, and raise the gain of the electron multiplier until the copper grid becomes visible on the secondary electron detector (SED). Select lens 1 and wobble. Use the multiple variable aperture (MVA) on the side of the instrument to improve the image quality. When best result is achieved stop wobble, select lens 2, and start wobble again. Adjust beam steering (x and y) to improve the image quality when needed. At optimal image quality stop wobble, select blanker, and start wobble again. This time adjust lens 2 to fix the image and lens 1 to refocus. After refocusing the procedure is repeated until a clear fixed and focused image of the copper grid can be seen on the SED. 2. In hardware make sure that there is no voltage on the bunching parameter (perform the imaging measurements in unbunched mode for optimal image quality). Before the start of the experiment select under acquisition setup/advanced settings: save as raw file, in order to be able to post-process the raw data after the measurement is completed. 3. Perform the ME-SIMS experiment in such a way that the analysis is conducted in the static SIMS regime. This can be achieved with a primary ion beam current of ∼450 pA, a primary pulse length of 30 ns, a spot diameter of 500 nm, and a primary ion energy of 15 kV. At 3 min per experiment this results in a primary ion dose of 4.9×1011 ions/cm2 (see Note 6). 4. For each chemical image, the primary beam is rastered over a 150×150 µm sample area, divided into 256×256 square pixels (larger or smaller areas can also be chosen). To image a significant larger surface, like a whole tissue section, analyze multiple 150×150 µm areas by stepping the sample stage in a mosaic pattern. To compensate for small deviations on the sample stage positioning a 10 µm overlap with the previous acquired sample is taken (the sample stage is moved by 140 µm). Each individual experiment is saved as .raw file to allow post-processing of the data. In “spectra” choose specific m/z ranges and select “image” for each range. Now in “acquisition”,“setup/advanced settings” select “acquire from raw file” and under the tab “image” the selected distributions can be seen. For large tissue sections two approaches of image stitching are available. First 150 µm2 images are stitched together manually in image handling software like Adobe Photoshop. Second a featurebased image alignment algorithm using PCA was developed for the visualization of the imaging data (28).

Cellular Imaging Using Matrix-Enhanced and Metal-Assisted SIMS

3.6. Discussion and Conclusions

205

Following the procedures discussed in Section 3 images were acquired as shown in Fig. 11.1. In panel A ME-SIMS images of lymnaea Stagnalis brain tissue are shown and in panel B MetA-SIMS images of single neuroblastoma cells are shown. The plasma coating procedure in MetA SIMS enhances desorption/ionization of membrane components such as lipids and sterols in SIMS imaging of tissues and cells. As shown

Fig. 11.1. (Panel 1) Direct molecular imaging of Lymnaea stagnalis nervous tissue by ME-SIMS. (a) Optical image of the Lymnaea cerebral ganglia, inset shows high magnification image of neurons in the anterior lobe (solid box), arrows indicate nuclei. Different regions in the section are right and left cerebral ganglia (Cgr and Cgl), anterior lobe (Al), commissure (Cm), and dorsal bodies (Db). (b) ME-SIMS image of APGWamide (429.0–433.2 m/z ; green, 0–3 counts) distribution. (c) ME-SIMS image of cholesterol (368.2–371.3 m/z ; blue, 0–4 counts) distribution. Scale bar: 200 µm; scale bar inset: 10 µm. Molecular images (b and c) are presented as colored overlays on top of the gray-scale TIC (total ion count) image (mass range: 1.0–5,000 m/z, 0–140 counts). The ME-SIMS measurements used indium primary ions (total ion dose 4.9 × 1011 ions/cm2 ). Reprinted from Ref. (10). (Panel 2) Cellular localization of MetA-SIMS selected ion count signals from neuroblastoma cells. Cells were imaged after deposition of 1-nm gold. (a) m/z 369 (cholesterol [M–OH]+ , 0–14 counts) and 607 (DAG, 0–6 counts). (b) m/z 970 (cholesterol [2 M+Au]+ , 0–4 counts) and 1,080 (0–1 counts). (c) m/z 369 (cholesterol [M–OH]+ , 0–12 counts) and 895 (0–2 counts). (d) m/z 607 (DAG, 0–4 counts) and 1,130 (0–1 counts). (e) m/z 848 (0–1 counts) and 895 (0–1 counts). Scale bar: 100 µm in panels (a–d) and 10 µm in panel E. Total ion dose 1.18 × 1012 ions/cm2 in panels (a–d) and 4.7 × 1012 ions/cm2 in panel (e). Reprinted from Ref. (27).

206

Altelaar and Piersma

in Fig. 11.1 High-resolution images of cholesterol and other membrane components can be obtained for single neuroblastoma cells and reveal subcellular details. Alternatively, in MESIMS, 2,5-dihydroxybenzoic acid electrosprayed on neuroblastoma cells and tissues allows intact molecular ion imaging of phosphatidylcholine (PC) and sphingomyelin (SM) at the cellular level.

4. Notes 1. Recent observations by Svensson et al. (29) point toward post-mortem changes in the proteome of susceptible peptides and proteins within minutes. To prevent these alterations of the sample Svensson et al. use focused microwave irradiation to sacrifice the animals. Since focused microwave irradiation is not available in every laboratory, an alternative way to minimizing alterations is snap-freezing of the dissected tissue and defrosting only just before sample preparation starts. 2. Be very careful not to make any contact between the TissueTek and parts of the tissue used for sectioning. The TissueTek contains a large amount of polymer substance, which smears over the tissue surface upon cutting. These polymers will dominate the resulting mass spectra. 3. After defrosting of the samples the degenerative processes in the sample continue by the reactivations of multiple proteases. For this reason the defrosting should be done just before sample preparation starts and tissues should not be left untreated for a sustained period of time. Furthermore, sample preparation should be done delicate but fast to prevent further degeneration and allow intact molecular species to be incorporated in the MALDI matrix. 4. The ESD needle can get clocked when too high matrix concentrations are used. Adjust the concentration of the matrix solution if needed. 5. Since in SIMS imaging experiments only the top layer is sampled, the matrix layer should not be too thick and the analytes have to be able to migrate to the top surface layer. This combination is critical for a successful ME-SIMS experiment and therefore the ESD conditions have to be optimized carefully, in order to produce a fine mist of matrix droplets so the matrix arrives onto the sample surface in a wet environment.

Cellular Imaging Using Matrix-Enhanced and Metal-Assisted SIMS

207

6. In static SIMS the same area is never sampled twice in order to prevent imaging of induced damage. To do so only 1% of the sample surface is analyzed, which converts to a primary ion dose of 1013 ions/cm2 . References 1. Vickerman, J. C., Briggs, D. (Eds.) (2001) ToF-SIMS: Surface Analysis by Mass Spectrometry, IM Publications and SurfaceSpectra Limited, Chichester. 2. Todd, P. J., McMahon, J. M., Short, R. T., McCandlish, C. A. (1997) Organic SIMS of biologic tissue. Anal Chem, 69, 529A–535A. 3. Delcorte, A., Garrison, B. J. (2000) High yield events of molecular emission induced by kiloelectronvolt particle bombardment. J Phys ChemB,104, 6785–6800. 4. Chandra, S., Morrison, G. H. (1995) Imaging ion and molecular-transport at subcellular resolution by secondary-ion massspectrometry. Int J Mass Spectrom Ion Process, 143, 161–176. 5. Chandra, S., Smith, D. R., Morrison, G. H. (2000) Subcellular imaging by dynamic SIMS ion microscopy. Anal Chem, 72, 104A–114A. 6. Strick, R., Strissel, P. L., Gavrilov, K., LeviSetti, R. (2001) Cation–chromatin binding as shown by ion microscopy is essential for the structural integrity of chromosomes. J Cell Biol, 155, 899–910. 7. Colliver, T. L., Brummel, C. L., Pacholski, M. L., Swanek, F. D., Ewing, A. G., Winograd, N. (1997) Atomic and molecular imaging at the single-cell level with TOF-SIMS. Anal Chem, 69, 2225–2231. 8. Pacholski, M. L., Cannon, D. M., Ewing, A. G., Winograd, N. (1998) Static time-of-flight secondary ion mass spectrometry imaging of freeze-fractured, frozen-hydrated biological membranes. Rapid Commun Mass Spectrom, 12, 1232. 9. Todd, P. J., Schaaff, T. G., Chaurand, P., Caprioli, R. M. (2001) Organic ion imaging of biological tissue with secondary ion mass spectrometry and matrix-assisted laser desorption/ionization. J Mass Spectrom 36, 355–369. 10. Altelaar, A. F. M., van Minnen, J., Jimenez, C. R., Heeren, R. M. A., Piersma, S. R. (2005) Direct molecular imaging of Lymnaea stagnalis nervous tissue at subcellular spatial resolution by mass spectrometry. Anal Chem, 77, 735–741. 11. Luxembourg, S. L., McDonnell, L. A., Duursma, M. C., Guo, X. H., Heeren, R.

12.

13.

14.

15.

16.

17.

18.

19.

20.

M. A. (2003) Effect of local matrix crystal variations in matrix-assisted ionization techniques for mass spectrometry. Anal Chem, 75, 2333–2341. McDonnell, L. A., Piersma, S. R., Altelaar, A. F. M., Mize, T. H., Luxembourg, S. L., Verhaert, P., van Minnen, J., Heeren, R. M. A. (2005) Subcellular imaging mass spectrometry of brain tissue. J Mass Spectrom, 40, 160–168. Wittmaack, K., Szymczak, W., Hoheisel, G., Tuszynski, W. (2000) Time-of-flight secondary ion mass spectrometry of matrixdiluted oligo- and polypeptides bombarded with slow and fast projectiles: positive and negative matrix and analyte ion yields, background signals, and sample aging. J Am Soc Mass Spectrom, 11, 553–563. Wu, K. J., Odom, R. W. (1996) Matrixenhanced secondary ion mass spectrometry: a method for molecular analysis of solid surfaces. Anal Chem, 68, 873–882. Nygren, H., Johansson, B. R., Malmberg, P. (2004) Bioimaging TOF-SIMS of tissues by gold ion bombardment of a silver-coated thin section. Microsc Res Tech, 65, 282–286. Adriaensen, L., Vangaever, F., Gijbels, R. (2004) Metal-assisted secondary ion mass spectrometry: influence of Ag and Au deposition on molecular ion yields. Anal Chem, 76, 6777–6785. Delcorte, A., Bertrand, P. (2004) Interest of Silver and Gold Metallization for Molecular SIMS and SIMS Imaging, 250–255, Elsevier Science, London. Delcorte, A., Bour, J., Aubriet, F., Muller, J. F., Bertrand, P. (2003) Sample metallization for performance improvement in desorption/ionization of kilodalton molecules: quantitative evaluation, imaging secondary ion MS, and laser ablation. Anal Chem, 75, 6875–6885. Delcorte, A., Medard, N., Bertrand, P. (2002) Organic secondary ion mass spectrometry: sensitivity enhancement by gold deposition. Anal Chem, 74, 4955–4968. Keune, K., Boon, J. J. (2004) Enhancement of the static SIMS secondary ion yields of lipid moieties by ultrathin gold coating of

208

21.

22. 23.

24.

25.

Altelaar and Piersma aged oil paint surfaces. Surf Interface Anal, 36, 1620–1628. Nguyen, T. C., Ward, D. W., Townes, J. A., White, A. K., Krantzman, K. D., Garrison, B. J. (2000) A theoretical investigation of the yield-to-damage enhancement with polyatomic projectiles in organic SIMS. J Phys Chem B, 104, 8221–8228. Sjovall, P., Lausmaa, J., Johansson, B. (2004) Mass spectrometric imaging of lipids in brain tissue. Anal Chem, 76, 4271–4278. Todd, P. J., McMahon, J. M., McCandlish, C. A. (2004) Secondary ion images of the developing rat brain. J Am Soc Mass Spectrom, 15, 1116–1122. Touboul, D., Halgand, F., Brunelle, A., Kersting, R., Tallarek, E., Hagenhoff, B., Laprevote, O. (2004) Tissue molecular ion imaging by gold cluster ion bombardment. Anal Chem, 76, 1550–1559. Townes, J. A., White, A. K., Wiggins, E. N., Krantzman, K. D., Garrison, B. J., Winograd, N. (1999) Mechanism for increased yield with SF5+ projectiles in organic SIMS:

26.

27.

28.

29.

the substrate effect. J Phys Chem A, 103, 4587–4589. Weibel, D., Wong, S., Lockyer, N., Blenkinsopp, P., Hill, R., Vickerman, J. C. (2003) A C-60 primary ion beam system for time of flight secondary ion mass spectrometry: its development and secondary ion yield characteristics. Anal Chem, 75, 1754–1764. Altelaar, A. F. M., Klinkert, I., Jalink, K., de Lange, R. P. J., Adan, R. A. H., Heeren, R. M. A., Piersma, S. R. (2006) Gold-enhanced biomolecular surface imaging of cells and tissue by SIMS and MALDI mass spectrometry. Anal Chem, 78, 734–742. Broersen, A., van Liere, R., Altelaar, A. F. M., Heeren, R. M. A., McDonnell, L. A. (2008) Automated, feature-based image alignment for high-resolution imaging mass spectrometry of large biological samples. J Am Soc Mass Spectrom, 19, 823–832. Svensson, M., Skold, K., Svenningsson, P., Andren, P. E. (2003) Peptidomics-based discovery of novel neuropeptides. J Proteome Res, 2, 213–219.

Chapter 12 Tandem Mass Spectrometric Methods for Phospholipid Analysis from Brain Tissue Timothy J. Garrett and Richard A. Yost Abstract We describe the utility of intermediate-pressure MALDI and tandem mass spectrometry (MS/MS and MSn ) for the characterization and imaging of phospholipids in brain tissue sections. The use of both MS/MS spectra and MS/MS images allows for identification of isobaric compounds. The structural characterization of phosphatidylcholines, phosphatidylserines, phosphatidylethanolamines, and sphingomyelins directly from tissue sections is described. Key words: MALDI, tandem mass spectrometry, phospholipids, brain.

1. Introduction Imaging tissue sections by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) is a growing field in mass spectrometry. This is due in part to the prospects of analyzing for very specific molecular ions (1–3) directly from sectioned tissue and correlating changes in those specific ions to diseases (4, 5). Critical to this comparison is the role of tandem mass spectrometry (MS/MS or MSn ) in the detection and identification of known and unknown compounds desorbed from the tissue (3, 6–9). The field of imaging mass spectrometry pertains to the direct analysis of surfaces, primarily tissue sections, using a focused ion beam (secondary ion mass spectrometry or SIMS) (10) or a focused laser beam with a highly absorbing matrix (MALDI) (11). In order to create an image, the laser or ion beam is rastered in a discrete pattern across the tissue surface. Each spot S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_12, © Springer Science+Business Media, LLC 2010

209

210

Garrett and Yost

that the laser interrogates generates a mass spectrum and is considered a pixel with dimensions related to the spot size of the laser or ion beam (12). A majority of the work in imaging MS has been focused on determining the distribution of proteins and peptides in tissue sections. Smaller molecules such as drugs (3, 8, 13, 14) and phospholipids (6, 7, 15–18) have also been analyzed, showing that a wide variety of compounds can be localized in tissue sections with MALDI-MS and identified by tandem MS (3, 6, 7, 9, 18). Phospholipid (PL) ions have been analyzed directly from tissue sections by mass spectrometry for several years. Secondary ion mass spectrometry (SIMS) offered images for the fragment ion at m/z 184, which corresponds to the head group for all phosphatidylcholines and sphingomyelins, showing distribution primarily in the gray matter of the brain (19). With the use of MALDI, researchers have been able to generate images of PLs from both molecular ions and fragment ions; tandem MS has clearly shown that several isobaric species are typically present at each m/z value in the lipid mass range (7, 9, 18). The direct analysis of PLs from brain tissue sections could aid in the further understanding of how lipids relate to diseases such as Alzheimer’s (20, 21) and Parkinson’s (22) by identifying specific lipids and the location of those lipids in a brain tissue section. Imaging PLs in brain tissue may also aid in understanding how and where PLs change with age and development (19, 23). Not only determining the molecular weight of PLs (by MS) but also identifying the fatty acid tails (by MS/MS) (6) may also aid in understanding if a decrease in certain fatty acids is PL specific or global (24) and if it is localized in a certain structure of the brain or other organ. Critical to this aspect of imaging MS is the need to identify compounds directly from the tissue section using tandem MS with fatty acid regiospecificity. Coupling a MALDI source to an instrument capable of performing tandem MS provides the opportunity to identify lipids based on characteristic fragmentation patterns and to produce images based on a specific fragmentation pathway (6) rather than on a molecular ion signal, giving a much more specific image related to a single, identified lipid. This specificity is particularly critical when analyzing small molecules, as both isobaric lipids and MALDI background ions can lead to multiple interferences; the reduction in noise provided by tandem mass spectrometry offers the ability to identify compounds present at very low levels that would not be detectable in a full-scan MS experiment. Acquiring tandem MS data for a specific parent ion across the tissue section can provide more specific images because a specific parent ion/product ion transition can be mapped. The image thus generated would be a tandem MS image. As has been shown previously (6), performing tandem MS on a ion trap provides the

Tandem Mass Spectrometric Methods

211

opportunity to separate and identify multiple isobaric PLs (as well as MALDI matrix background ions) from the isolation of parent ions of a single m/z. Tandem MS with MALDI thus offers the ability to completely characterize the compounds desorbed and ionized from a tissue section and differentiate them from those related to the MALDI matrix. The methods presented here describe the ability to characterize phosphatidylcholines (PCs), sphingomyelins (SPMs), phosphatidylserines (PSs), and phosphatidylethanolamines (PEs) in the positive ion mode from rat brain tissue sections using imaging tandem MS and the creation of compound-specific images from full-scan MS and MS2 , while using MS3 for the final identification. The use of MS2 images for the separation of isobaric ions coupled with the ability to perform MS3 provides the means to identify isobaric species present in the tissue section.

2. Materials 2.1. Equipment

1. Artistic airbrush (Aztek A470) from Testors (Rockford, IL, USA). 2. Thermo LTQ with vMALDI ion source (see Note 1). 3. Cryostat (Leica CM1850, Wetzlar, Germany). 4. Surfer 8.0 or newer version (Golden Software, Golden, CO, USA).

2.2. Reagents and Supplies

1. 40 mg/ml solution of 2,5-dihydroxybenzoic acid (DHB) prepared in 70/30 (v/v) HPLC-grade methanol and water. Sodium acetate is added to a concentration of 20 mM for sodium adduction, if desired. 2. Double-sided tape for attaching microscope slides to MALDI stage. 3. Glass microscope slides (see Note 2).

3. Methods 3.1. Preparation of Brain Tissue Sections

1. Whole brain tissue (from Sprague–Dawley rats in this example) should be cut with a scalpel to desired location of sectioning. ◦

2. Set temperature of cryostat to –22 C.

212

Garrett and Yost

3. Affix tissue to cutting target by holding tissue in one hand and pipetting a small amount of water around the tissue. 4. Set cutting thickness to 10 µm. 5. Cut tissue several times until an entire section is obtained. Discard unwanted sections. 6. When desired section is sliced, attach to a cold glass microscope slide by placing the slide over top of the section. Press down lightly on the section for it to adhere to the slide (see Note 3). 7. Remove slide with tissue from cryostat chamber. With a gloved finger, warm the underneath side of the glass below the tissue until the tissue becomes clear. This takes approximately 10 s. 8. Freeze until ready for use. 3.2. Application of MALDI Matrix with an Artistic Airbrush

1. Remove desired tissue sections from the freezer and place in a dessicator for at least 30 min. 2. Prepare a solution of DHB to a concentration of 40 mg/ml in 70/30 methanol/water with 20 mM NaCH3 COO (see Note 4). 3. Set up airbrush in a ventilated hood. 4. Attach the airbrush to a ring stand using an appropriate clamp. 5. Adjust the distance from the tissue to the end of airbrush to be approximately 15 cm (see Note 5). 6. Attach a second clamp to the solvent and gas control throttle. Ensure that the trigger is clamped down and that solvent is flowing. 7. Place the blank microscope slide underneath the airbrush setup and pass underneath the spray in a back and forth motion at a steady pace. 8. After a couple of passes have been made, examine the glass slide under light to evaluate the coating evenness. A gloved finger can be used to test the adherence of the matrix by wiping the surface of the glass slide. If smearing is evident, the coating has been applied too wet, if the matrix comes off as a powder, the coating was applied without enough solvent. 9. After determining the appropriate settings, obtain the tissue sample and pass it underneath the airbrush assembly in a steady back and forth motion as before. Allow for a little bit of overlap as you move across the tissue and be sure to coat beyond the edges of the tissue. One pass is considered

Tandem Mass Spectrometric Methods

213

a full covering of the tissue. Typically, 20 passes are needed to achieve an appropriate coating. 10. Dry the completed sample for a minimum of 30 min under a gentle stream of nitrogen. 11. Place the sample on the MALDI target plate into the predrilled wells and attach with double-sided tape. 3.3. Mass Spectrometric Analysis

1. Insert the MALDI target plate with the tissue into the instrument following the manufacturer’s procedure. 2. For the instrument used here, a picture of the tissue sample needs to be obtained first. This is done by selecting the area where the tissue is located and then pressing scan to obtain the image. 3. Select the area on the tissue for analysis. Encircle the entire tissue sample using a desired geometric shape or free draw application if the whole tissue is desired for analysis (see Note 6). 4. Set the desired mass range; for PL analysis, we typically use the mass range of 180–1,000. 5. If automatic gain control (AGC) is not used, the laser power and the number of laser shots per spot must be set manually. This is done by first choosing the desired number of shots (typically 5 or 10) and then adjusting the power of the laser to achieve a signal of 1×105 . A minimum of 10 random spots should be interrogated to determine the appropriate power. If employing AGC, the power only needs to be adjusted to obtain a similar signal as the number of laser shots are varied when AGC is employed (see Note 7). 6. Set the raster size; typically, this is set to the spot size of the laser for normal sampling (see Note 8).

3.4. Tandem MS Analysis

1. The isolation width is typically set to 1.5 amu with a collision energy of 30% (see Note 9). 2. For obtaining tandem MS images, the entire tissue section must be analyzed with the given m/z of interest. This provides for the identification of potential isobaric ions as well as the generation of MS/MS images corresponding to selected product ions. 3. If tandem MS will only be used for compound identification, a second procedure could be used. Following the instrument manufacturer’s procedures, a desired list of ions can be entered. In this mode, the entire tissue will be analyzed alternating between the list of ions entered during the tissue scanning.

214

Garrett and Yost

3.5. Image Generation

1. MS images from full-scan MS and MS/MS data are generated by extracting a desired m/z value with a 1.0 m/z wide window using a MALDI data extraction program provided by ThermoFisher. The extracted data is in text form and shows the x, y coordinates and the intensity of the ion. 2. The text version of the data is converted to a spreadsheet using a spreadsheet program (Microsoft Excel). 3. A graphical program is used to generate ion images from the spreadsheet (Surfer 8.0).

3.6. Results 3.6.1. Identifying the Lipid Signature

We have previously shown the ability to analyze phospholipids in tissue sections by intermediate-pressure MALDI (67); this method describes a comprehensive procedure for the evaluation of those phospholipids. Figure 12.1 shows an optical image of a brain section (inset) and a mass spectrum from one point in that section. The circle and arrow in the figure indicate the area from which this mass spectrum was acquired (1 scan, 10 laser shots, scanned from m/z 150–1,000). The majority of the ion signal falls between m/z 700 and 900. From previous results (9, 15), the ions from m/z 650–900 can be identified as arising from PLs; searching a database of PL

Fig. 12.1. The inset is an optical image (9×15 mm) generated from inside the mass spectrometer of a rat brain section coated with DHB matrix. It was acquired with 1×1 mm square pictures that are stitched together; this creates the lines in the picture. The mass spectrum shows the signal from the area on the tissue indicated by the circle and arrow. The open circle is approximately equal to the laser spot size (120 µm).The spectrum was acquired with 10 laser shots. A total of 11,156 spectra were collected across the tissue section.

Tandem Mass Spectrometric Methods

215

ions (constructed in the lab) generated a large list of possible identities for each ion based only on the m/z value submitted. The list included PCs, SPMs, PEs, and PSs as either protonated, [M+H]+ , or sodiated, [M+Na]+ , ions. The ion at m/z 184 results from source fragmentation of only protonated PCs and SPMs; it corresponds to the phosphocholine head group from these two groups of PLs (25). The low intensity of the m/z 184 fragment ion is primarily due to the fact that sodium acetate was added to the matrix solution to induce the production of primarily cationized adduct ions ([M+Na]+ ). Producing the sodiated adduct has three advantages: (1) the intensity of the fragment ion at m/z 184 is significantly reduced, (2) the sodiated species are a more stable ion for mass analysis (26), and (3) the fragmentation patterns for collision-induced dissociation (CID) of cationized PLs show more structural information (25, 27, 28). In addition, the intensity of the m/z 184 fragment ion is also reduced by the use of intermediate pressure in the MALDI source (6). Table 12.1 shows the identification of 16 ions in the PL region from an experiment in which the tissue section was

Table 12.1 List of compounds identified from a brain tissue section using tandem MS with an imported mass list m/z

RA (%)

PL

Ion

694.9

8.7

SPM (18:0)

[M+Na–N(CH3 )3 ]+

697.8

9.4

PC (16:0, 16:0)

[M+Na–N(CH3 )3 ]+

723.8

16.0

PC (16:0, 18:1)

[M+Na–N(CH3 )3 ]+

753.9

34.7

SPM (18:0)

[M+Na]+

756.8

59.8

PC (16:0, 16:0)

[M+Na]+

772.7

8.6

PC (16:0, 16:0)

[M+K]+

782.7

100.0

PC (16:0, 18:1)

[M+Na]+

798.6

33.8

PC (16:0, 18:1)

[M+K]+

804.7

22.3

PC (16:0, 20:4)

[M+Na]+

810.7

30.3

PC (18:0, 18:1)

[M+Na]+

828.6

16.4

PC (16:0, 22:6)

[M+Na]+

832.6

17.2

PC (18:0, 20:4)

[M+Na]+

848.6

7.1

PC (18:0, 20:4)

[M+K]+

856.7

7.0

PC (18:0, 22:6)

[M+Na]+

932.4

4.4

PC (16:0, 16:0)

[M+DHB+Na]+

958.4

5.8

PC (16:0, 18:1)

[M+DHB+Na]+

A single compound such as PC (16:0, 18:1) can be represented by four different ions in the mass spectrum, m/z 723.8 ([M+Na–N(CH3 )3 ]+ ), 782.7 ([M+Na]+ , 798.6 ([M+K]+ ), and 958.4 ([M+Na+(DHB+Na−H)]+ ). The abbreviation RA stands for relative abundance

216

Garrett and Yost

scanned and MS/MS spectra were recorded for each ion in a list. In this type of experiment, a list of ions is imported into the method. When using one microscan per spot, the instrument rotates through the list while rastering across the tissue section. Thus, with a list of 16 ions (as was done here), a single ion is analyzed every 16 spots on the tissue. This experiment provided a way to collect MS/MS data for multiple ions in one tissue raster; however, MS/MS images could not be generated. As can be seen in Table 12.1, only two classes of PLs were identified using this approach, SPM and PC. From these studies, it was determined that a single PC ion can have multiple ions associated with a single fatty acid arrangement. For example, PC (16:0, 18:1) has four ions that were detected, as shown in Table 12.1, [M+Na–N(CH3 )3 ]+ , [M+Na]+ , [M+K]+ , and [M+DHB+Na]+ , as determined by tandem MS. In addition, there is also a low intensity [M+H]+ ion present in the spectrum (m/z 760.8, ∼8% RA). MS/MS was performed on this ion (in a separate experiment) and the major fragment produced was m/z 184.1, confirming the protonation. Tandem MS (MS/MS) allowed for the identification of these 16 ions from a single tissue scan; however, only two lipid classes were identified from this type of experiment in which selected ions were chosen for tandem MS, while scanning the tissue section. In addition, several SPMs were not characterized, SPM 24:0 (m/z 837.6) and SPM 24:1 (m/z 835.6) (both [M+Na]+ ), by this method because they were not chosen for MS/MS analysis. These two ions were not selected because they were not very intense in a visual inspection of the spectrum. However, these two ions should only be present in the white matter of the brain according to previous research (29). The spectrum used to select ions for subsequent MS/MS experiments probably did not include as much white matter as gray matter; therefore, the intensity of these two ions would be low and would not have been selected for MS/MS. However, because a full-scan MS data set was also collected, the image for any ion can be retrieved by extracting the specific ion signals. From previous work (6 7), the most abundant ion for all the PC and SPM ions under the conditions employed here is the [M+Na]+ . Images generated for these two ions confirmed their greater abundance in white matter. 3.6.2. Imaging with Tandem MS Data

In order to further characterize the PLs and generate MS/MS images, another tissue section was prepared for imaging MS experiments. In several previous experiments, it was noticed that a single tissue section could be analyzed multiple times for the analysis of PLs (up to 40 times) (6); therefore, to explore the further characterization of the PL’s MS/MS and MS3 experiments were performed in which a single m/z value was isolated (typically with an isolation width of 1.5) and the entire tissue was

Tandem Mass Spectrometric Methods

217

analyzed by MS/MS or MS3 for that specified ion and isolation width. Table 12.2 shows the experiments that were performed on this tissue section and the ions isolated for tandem MS. This list represents 45 different experiments, but in total, 49 experiments were performed, including four full-scan MS experiments. Two more MS/MS experiments and two full-scan experiments were performed (after recoating the tissue section because the signal was depleted after 41 experiments). Figure 12.2 shows the ion images from a full-scan MS experiment for the same 16 PL species identified earlier (Table 12.1),

Table 12.2 List of tandem MS experiments performed on a single tissue section MS2 (IW 1.5)

MS3

742.6 → . . .

753.6 → 694.4 → . . .

753.6 → . . .

756.6 → 697.3 → . . .

756.6 → . . .

804.6 → 745.4 → . . .

782.6 → . . .

808.6 → 552.3 → . . .

798.6 → . . .

808.6 → 749 → . . .

804.6 → . . .

808.6 → 765 → . . .

808.6 → . . .

826.6 → 767 → . . .

810.6 → . . .

828.6 → 741.3 → . . .

820.6 → . . .

828.6 → 769.4 → . . .

826.6 → . . .

828.6 → 785.2 → . . .

828.6 → . . .

832.6 → 745 → . . .

832.6 → . . .

832.6 → 773 → . . .

835.6 → . . .

835.6 → 776 → . . .

835.6 (IW 1.0) → . . .

835.6 → 793 → . . .

836.6 (IW 1.0) → . . .

836.6 → 777 → . . .

837.6 → . . .

837.6 → 778 → . . .

848.4 → . . .

837.6 → 794 → . . .

856.6 → . . .

856.6 → 769 → . . . 810.6 → 751 → . . . 826.6 → 761 → . . . 848.6 → 789.4 → . . . 856.6 → 797.4 → . . .

Those in italics were performed after recoating the tissue. In addition to these experiments, two full-scan MS analyses were performed before recoating and after recoating. In total, 49 experiments were performed on the same tissue section for the analysis of phospholipids. More experiments could have been performed after recoating, but the data collected were sufficient to identify the ions of interest. The abbreviation IW stands for isolation width

218

Garrett and Yost

but from a different tissue section. A full-scan MS experiment is critical to perform for endogenous species analysis because it provides a starting point for knowing what major ions are present in the tissue section. These images were generated by extracting the intensity of each specified m/z with a window of 1.0 amu. These ions were all identified by tandem MS, as indicated in Table 12.2. As can be seen from this set of 16 ion images, the different PLs (PC or SPM) show different distributions in the rat brain. It is remarkable that the change of a single fatty acid tail, such as with m/z 782 (PC (16:0, 18:1)) and m/z 810 (PC (18:0, 18:1)), can so drastically change the distribution in the rat brain. These MS images offer a unique perspective of the rat brain, showing that the distribution of individual PL ions is varied throughout the entire brain section. Not surprisingly, m/z 184, the fragment ion that corresponds to the polar head group of all PC and SPM ions, shows a distribution similar to that of m/z 782, the most abundant PC ion detected in the positive ion mode. What is important is that imaging just the fragment ion at m/z 184 as has been performed in other imaging MS studies (19) does not provide enough information about the localization of other PC and SPM

Fig. 12.2. MS images from one data file for 16 phospholipid species. See Table 12.2 for the identification of each ion. The maximum value of the intensity scale is indicated at the bottom left of each image. All images were normalized to the total ion current.

Tandem Mass Spectrometric Methods

219

ions and thus provides a very limited view of the chemical distribution of the rat brain. Since MS/MS was performed across the entire tissue on 14 of the 16 ions in Fig. 12.2, MS/MS images were also generated from the most abundant loss produced in each MS/MS experiment. The most abundant loss for all ions isolated was a neutral loss of 59, corresponding to a loss of trimethylamine from either a PC or a SPM ion. The MS/MS ion images for 14 of these ions are shown in Fig. 12.3. The ion at m/z 798 was determined from previous results to be the [M+K]+ ion for PC (16:0, 18:1), corresponding to the [M+Na]+ ion of PC (16:0, 18:1) at m/z 782. A full tissue scan MS/MS experiment was not performed on the fragment ion at m/z 184 nor on the ion at m/z 798. The MS/MS transition for each ion is shown above each MS/MS ion image. However, performing an MS/MS experiment for m/z 798 might have been beneficial to determine the presence of other isobaric ions, as will be described later. For the most part, the MS/MS images

Fig. 12.3. MS/MS images for only 14 different PC and SPM species. The neutral loss of 59 for each ion is shown. See Table 12.2 for the identification of each ion. Some ions presented represent different ions of the same compound such as 810→751 and 826→767 ([M+Na]+ and [M+K]+ for PC 18:0 and 18:1, respectively). These signals were not normalized to the total ion current because they are tandem MS images. Maximum scale intensity is shown in the bottom left of each figure.

220

Garrett and Yost

are similar to the corresponding MS images. However, for some ions, this more specific image provides a different view of the ion distribution in the tissue. For example, the MS image of m/z 810 indicates that this ion is present in both gray and white matter of the brain and thus shows a distribution that is fairly uniform throughout the tissue. On the other hand, the MS/MS image of 810–751 shows that the distribution of the PC (18:0, 18:1) [M+Na]+ ion is predominantly in the white matter of the brain. This means that the MS image of m/z 810 included other species at a similar intensity and of the same nominal m/z, but predominantly in the gray matter. For m/z 836, the difference between the MS and MS/MS images is striking as well. The MS image for m/z 836 shows a uniform localization in the white matter of the brain, while the MS/MS image for m/z 836 → 777 depicts localization to the substantia nigra and red nucleus, with lesser amounts in the corpus collosum of the brain. In the case of the less abundant ions (m/z 826, 835, 837, and 848), the MS/MS data allowed for a much clearer picture of the distribution than the MS image. This clarity is likely due to the improvement in the signal-to-noise ratio afforded by tandem MS and the selectivity of monitoring a single MS/MS daughter ion. In all cases, the MS/MS image should provide a better representation (compared to just MS alone) of an individual compound, because interfering isobaric species are removed if they fragment differently. The ability to separate isobaric ions using tandem MS in imaging tissue has been shown previously (6), but the characterization of the many isobaric ions present has not been fully explored. For example, Cha and Yeung (18) showed the characterization of cerebrosides and PCs using graphite-assisted laser desorption and the separation of isobars including cerebrosides, but not across the entire tissue. In the current study, tandem MS data were collected across the entire tissue from over 40 different ions. The full-scan tandem mass spectra were then analyzed for possible isobaric species present; see Table 12.2 for a list of all the tandem MS experiments. In addition, to help characterize the isobars, MS3 was performed on selected MS/MS daughter ions to elucidate the structure of the PL species present (see Table 12.2). 3.6.3. Isobaric Ion Identification

The process for characterization of the isobars from the ion signal at a single m/z is illustrated in Figs. 12.4 and 12.5. Figure 12.4 shows the MS2 spectrum of m/z 856 averaging all 10,528 spectra across the tissue section. The most abundant product ion, m/z 797.4, arises from a neutral loss of 59, corresponding to a loss of trimethylamine, and thus identifies one of the isobaric ions as a PC (not an SPM because of the even m/z at 856). Other relatively abundant and characteristic product ions include m/z 701.9, 769.4, 813.0, and 838.1, arising from neutral losses of 155, 87, 43, and 18, respectively. A neutral loss of 155 is typically

Tandem Mass Spectrometric Methods

221

m/z m/z

m/z

m/z

m/z

m/z

Fig. 12.4. Identification of the isobaric ions present at m/z 856.6 using MS/MS data. The MS/MS spectrum for m/z 856.6 (IW 1.5) is the average of 10,528 Spectra across the entire tissue section. Three major ions were identified at this m/z value and isolation width. The MS/MS images for the major fragment ion of each compound are shown in the left of the figure. The two insets in the spectrum show expanded regions as indicated. The fragment ions corresponding to each isobar are identified with colors and symbols. Axes are in micrometers.

related to a DHB cluster ion (loss of C7 H7 O4 ), a neutral loss of 43 is related to a PE ion (loss of C2 H5 N), and a neutral loss of 87 is typically associated with a PS compound (loss of C3 H5 NO2 ). The tandem MS images for m/z 856 dissociating to form m/z 797.2 (the PC isobar, NL of 59), 701.9 (DHB cluster, NL 155), and m/z 769.4 (the PS isobar NL 87) are shown at the left in Fig. 12.4. These data provide additional confirmation that there are multiple ionic species present at m/z 856.6, given that m/z 797.4 and m/z 769.4 show a near opposite distribution, as has been shown for other m/z values (6, 18). In addition, the MS image for the distribution of the DHB cluster shows a higher localization off the tissue, helping to confirm that this ion signal is from the matrix and not from the tissue section. The two inset spectral regions were expanded to show the importance of these less abundant ions in the identification of the isobars. Colors and symbols indicate the fragment ions that are representative of the different isobars present. Another issue with characterizing the compounds desorbed from tissue is also determining what type of ion is detected

222

Garrett and Yost

Fig. 12.5. MS/MS and MS3 data showing the identification of one of the isobars at m/z 856. A is the MS/MS spectrum from tissue for m/z 856.6. (a) NL of 59 is indicative of a PC or SPM, while a NL of 87 or 43 is indicative of PS or PE, respectively. (a) NL of 154 or 155 corresponds to a DHB cluster ion. (b) The average MS3 spectrum for 856→769→. . .. For this MS3 experiment, the entire tissue was scanned and thus the spectrum is the average of 10,528 spectra. The fragment ion at m/z 769 was identified as resulting from a PS ion because of the NL of 87. MS3 spectra from tissue (b) and a PS standard mixture from bovine brain (c). The assignment of the fatty acid chains is R1 =stearic acid (18:0) and R2 =oleic acid (18:1). The TIC values shown correspond to the signal from the standard (a) and from the tissue (b). MS3 was required for the correct assignment of the fatty acid tails of this isobaric species and identifies this PS species as PS (18:0, 18:1) [M–2H+3Na]+ .

(protonated, sodiated, etc.). The further characterization of the PS ion can be used to illustrate this point. The middle spectrum (b) of Fig. 12.5 is the MS3 spectrum acquired from the tissue section of the ion at m/z 769 arising from a neutral loss of 87 (856→769→. . .), typically associated with a PS compound (loss of C3 H5 NO2 , serine head group) (30). As can be seen from the figure, this MS3 fragmentation pattern shows four major product ions (m/z 505.3, 485.3, 465.3, and 463.3) in four very distinct patterns. A standard mixture of PSs (bovine brain extract from Avanti lipids) was obtained to evaluate the ionization and fragmentation characteristics of this class of lipids under the conditions employed. The two most abundant fatty acid arrangements in the mixture are PS (18:0, 18:1) and PS (18:0, 22:6). MALDI was performed on this standard mixture with DHB as the matrix. For the most abundant PS compound, PS (18:0, 18:1), multiple ions of this single compound were observed in the mass spectrum, including m/z 790.3 ([M+H]+ ), m/z 812.4 ([M+Na]+ ), m/z 834.4 ([M−H+2Na]+ , and m/z 856.4 ([M–2H+3Na]+ ).

Tandem Mass Spectrometric Methods

223

The next most abundant PS compound observed in the standard mixture is PS (18:0, 22:6) with three ions detected at m/z 858.4 ([M+Na]+ ), m/z 880.3 ([M−H+2Na]+ ), and m/z 902.7 ([M–2H+3Na]+ ). The four ions of the PS (18:0, 18:1) were each subjected to MS/MS and MS3 experiments. All ions containing a sodium show the most abundant fragment ion as a neutral loss of 87 (the serine head group, C3 H5 NO2 ) (data not shown); thus, this first stage of tandem MS can rule out that the ion is an [M+H]+ ion. MS3 was then performed on the ions arising from the NL of 87 from each of the three sodium adducts. Each MS3 spectrum produced a different fragmentation pattern, which was used to identify the corresponding ion from the tissue tandem MS experiment. Figure 12.5c shows the MS3 spectra of 856→769→. . . of the [M–2H+3Na]+ ion from the standard. The identification of each product ion is shown in the top spectrum. The most abundant ion, m/z 465.3, corresponds to the loss of R2 CO2 Na; the related loss of the fatty acid at R1 occurs at m/z 463.3 (loss of R1 CO2 Na). The MS3 spectra from the standard (C) and the tissue (B) are nearly identical, thus concluding that one of the isobaric ions at m/z 856.6 from the tissue is PS (18:0, 18:1) [M–2H+3Na]+ . With a better idea of how to identify possible isobars at a given m/z value, a better characterization of the ions detected can be accomplished. The identification of the PLs for the tissue section analyzed is described in Table 12.3. As can be seen, there is a DHB cluster ion detected at every mass isolated (typically arising from a neutral loss of 137, 154, 155, or 176). Many of the ions isolated for tandem MS show three and even four lipid species. The three primary lipid classes detected in the positive ion mode were PC, PS, and PE. Cerebrosides were also detected, but were present at a much lower intensity due to the conditions employed (18). The primary reason for maintaining an isolation width of 1.5 for most experiments was that PLs are considered fragile ions in ion trap analysis (26, 31) and thus can be lost during isolation if the isolation width is not wide enough. It should be noted that the isolation width was decreased from 1.5 to 1.0 for the classification of 835.6 and 836.6 because of confusion about a fragment ion observed, as detailed below. When averaging all the MS/MS data collected from across the tissue for m/z 835.6 with an isolation width of 1.5, the spectrum included a very intense ion at m/z 552.3. This corresponds to a neutral loss of 283, which would indicate an SPM (because the parent mass is odd) and the loss of the amide-linked fatty acid (18:0). However, most SPM ions do not produce an abundant neutral loss indicating the loss of an amide-linked fatty acid tail in MS/MS experiments. With an isolation width of 1.5 for a parent ion of m/z 835.6, this product ion could arise from ions with m/z 0.75 units higher or lower. The tissue was then rescanned in

PC (16:0, 22:6)

828

PC (18:1, 20:1)

SPM (24:0)

PC (18:0, 20:4)

PC (18:0, 22:6)

836

837

848

856

PC (18:0, 20:4)

PC (18:0, 18:1)

826

SPM (24:1)

PC (16:0, 20:4)

820

832

PC (18:0, 18:1)

810

835

PC (16:0, 20:4)

PC (18:1, 18:1)

804

808

PC (16:0, 18:1)

PC (16:0, 18:1)

782

798

SPM (18:1)

PC (16:0, 16:0)

753

PC (31:0)

742

756

PC or SPM ion

m/z

PS (16:0, 18:1)

[M−2H+3Na]+

DHB cluster

PS (18:0, 18:1)

DHB cluster

[M+Na]+

[M−2H+3Na]+

DHB cluster

[M+K]+

DHB cluster

PE (18:0, 22:6)

[M+Na]+

DHB cluster [M−H+2Na]+

[M+Na]+

DHB cluster

PE (18:0, 22:5)

[M+Na]+

PS (18:1, 18:1)

DHB cluster [M−H+2Na]+

PE (18:0, 20:4)

[M+Na]+

[M+K]+

DHB cluster

[M+Na]+

[M−H+Na+K]+

DHB cluster

DHB cluster

[M+K]+

[M−H+2Na]+

DHB cluster

[M+K]+

PE (18:1, 20:4)

[M+Na]+

[M−H+2Na]+

DHB cluster

PE (18:1, 18:1) PE (16:0, 22:6)

DHB cluster

[M+Na]+

DHB cluster

[M+K]+

[M+Na]+

DHB cluster

DHB cluster

Others

[M+Na]+ [M−H+2Na]+

Ion detected

DHB cluster

[M+Na]+

PS ion

[M+Na]+

PE (34:0)

[M+Na]+

Ion detected

[M+Na]+

PE ion

Ion detected

Table 12.3 List of ions identified by imaging tandem MS (MS2 full tissue scan). MS3 was performed for further identification when needed

224 Garrett and Yost

Tandem Mass Spectrometric Methods

225

MS/MS for m/z 835.6 with a narrower isolation width (1.0); the intensity of m/z 552.3 was significantly reduced. To identify the ion one m/z unit higher, the tissue was scanned for m/z 836.6 with an isolation width of 1.0. Fragments arising from the neutral loss of 43 (m/z 793.3), 59 (m/z 777.4), and 87 (m/z 749.4) were all detected, as well the ion at m/z 552.3 (neutral loss of 284). The major ion is assigned to PC (18:1, 20:1) [M+Na]+ , as signified by the neutral loss of 59; the PE ion was identified as PE (18:0, 22:6) [M−H+2Na]+ , giving rise to the neutral loss of 43. Finally, the ion related to PS (NL of 87) could be PS (40:6) [M+H]+ , but is not likely because a loss of 87 is less intense for protonated PS ions than for sodiated PS ions. MS3 was not performed on m/z 836.6→749.4, but could be used to help identify this isobar. 3.6.4. Understanding Endogenous Interferences

In the analysis of full-scan MS data from this tissue sample, an interesting observation was made. An ion, m/z 616, appeared to be localized within a specific region of the rat brain section (Fig. 12.6 inset). MS/MS was performed on this ion, and the most abundant signal was m/z 557.2, corresponding to a neutral loss of 59 (Fig. 12.6a). Coupling this result to those obtained from the analysis of all ions up to this point would suggest that

Fig. 12.6. MS/MS through MS5 spectra of m/z 616 from a tissue-specific region in a brain tissue section (shown in the inset). Even after MS5 , the compound still remained unidentified.

226

Garrett and Yost

this ion is a PC (even parent mass and neutral loss of 59). However, most PC ions are greater than m/z 700 when singly charged and thus one might conclude that this ion could represent a lysophosphatidylcholine (a PC ion that has lost one fatty acid tail). However, when MS3 was performed on m/z 557.2, a second neutral loss of 59 was produced, which is not consistent (or possible) for a PC ion. Additional tandem MS spectra were collected up to MS5 , using the most abundant ion from each stage, to assist in the identification of the compound (Fig. 12.6b–d). In a personal communication, this ion was suggested to be heme (M+ ). A heme standard was obtained and spotted onto a MALDI plate with α-Cyano-4-hydroxycinnamic acid (α-cyano). The most abundant ion signal detected was m/z 616. Tandem MS was performed up to MS5 (Fig. 12.7a–d), and the tandem MS spectra show nearly identical fragmentation patterns of the ion desorbed from the brain tissue, confirming the identification of m/z 616 as heme.

Fig. 12.7. MS/MS through MS5 spectra from a standard of heme (m/z 616), showing a good match in fragmentation pathways when compared to the MSn spectra acquired from the tissue section shown in Fig. 12.6.

3.7. Conclusions

The mass spectrometric images of the various PLs in rat brain showed a remarkable variation in distribution. The use of imaging tandem MS provided a means to properly identify the ions detected and to show the many isobaric species present in the lipid mass region. In addition, it was determined that there were sometimes up to five different ions found under MALDI conditions

Tandem Mass Spectrometric Methods

227

and characterized by tandem MS that represent the same compound, adding more complexity to the direct analysis of tissue for PLs. These different ions include protonated, sodiated, and potassiated species, as well as ions that are adducted with DHB matrix molecules and even ions with multiple cation adductions. Additionally, MS5 was shown to properly identify the unexpected compound heme in the tissue. Because heme loses 59 in MS/MS, a single stage of tandem MS might identify this compound as a lysophosphatidylcholine, which would be incorrect. This finding further shows the need of tandem MS in the analysis of small molecules from tissue sections and the need for multiple stages of tandem MS to correctly identify small molecules from tissue. Most of the images presented were generated from a single mass spectral acquisition, showing the wealth of information hidden within the data. The repeated analysis (MS, MS/MS, and MS3 ) of one individual brain section was described and offers a unique opportunity to continually probe a single tissue section for further information. Images generated from an MS/MS product ion in tandem MS experiments showed better specificity and in some cases a more accurate representation of the distribution of the compound when compared to the MS ion image.

4. Notes 1. The Finnigan LTQ linear ion trap mass spectrometer equipped with the vMALDI source (ThermoFisher, San Jose, CA, USA) was used for all imaging mass spectrometry experiments. The instrument has been described in detail previously (6), but briefly, the source consists of a N2 laser (337 nm) directed to the vacuum chamber using fiber optics. The laser spot size for these experiments was 120 µm. The spot size was evaluated by ablating a small piece of photosensitive paper and then measuring the size of the ablated area with a microscope. The source is maintained at a pressure of 0.17 Torr using N2 . 2. Even though non-conductive sample surfaces are used, the mass accuracy and peak shape are very good. Surface charging effects when using non-conductive glass microscope slides have been observed in MALDI time-of-flight imaging experiments (4), but are eliminated by the use of intermediate-pressure MALDI and a mass analyzer such as the linear ion trap that does not rely on the initial kinetic energy of the ions (6).

228

Garrett and Yost

3. It is important for the microscope slides to be cold before the tissue is applied to them; otherwise, the tissue will thaw instantaneously and will not likely be flat. The glass slides can be placed in the chamber of the cryostat during the cutting process. 4. Addition of sodium was employed to induce cationization over protonation because cationization provided for better structural characterization of the phospholipids studied (6). 5. A small surface to lay the microscope slide upon can be used to move the microscope slide underneath the airbrush setup. We use a piece of Plexiglas that is approximately 1.3-cm thick and measures approximately 15×15 cm. 6. Although the brain tissue sections are not square, a square was chosen to select the area of interest. This allowed for the analysis of regions outside the brain tissue section as a reference to the ion signals obtained from the tissue section. It also provides for evaluation of possible analyte migration off the tissue. 7. Automatic gain control (AGC) was turned off in order to maintain the same number of laser shots at each point. To avoid deleterious space-charge effects, the number of laser shots and the power of the laser were first determined by interrogating one spot of the tissue sample. Typically, 10 laser shots with about 20–30% laser power (arbitrary units) were used for all full-scan analyses and 15–20 laser shots were used for MS/MS and MS3 experiments. 8. The raster step size can be set to any desired value, but in the case of these experiments the step size was set to 120 µm, the same size of the laser spot size, to avoid over-sampling and to allow for repeated analysis of the same tissue section (6). 9. An isolation width of 1.5 amu is typically employed for PL analysis because the ions tend to be fragile in ion trap mass analysis (26). In some cases, a width of 1 amu is employed to remove potential background interferences.

Acknowledgments The authors wish to acknowledge the assistance of George Stafford and Mari Prieto-Conaway at ThermoFisher Scientific (San Jose, CA, USA). We thank Dr. Nigel Calcutt of the

Tandem Mass Spectrometric Methods

229

University of California San Diego for donation of the tissue specimens. The NIH is greatly acknowledged for support (NIH RO1 ES007355). References 1. Monroe, E. B., Jurchen, J. C., Lee, J., Rubakhin, S. S., Sweedler, J. V. (2005) Vitamin E imaging and localization in the neuronal membrane. J Am Chem Soc, 127, 12152–12153. 2. Hsieh, Y., Casale, R., Fukuda, E., Chen, J. W., Knemeyer, I., Wingate, J., Morrison, R., Korfmacher, W. (2006) Matrix-assisted laser desorption/ionization imaging mass spectrometry for direct measurement of clozapine in rat brain tissue. Rapid Commun Mass Spectrom, 20, 965–972. 3. Drexler, D. M., Garrett, T. J., Cantone, J. L., Diters, R. W., Mitroka, J. G., PrietoConaway, M. C., Adams, S. P., Yost, R. A., Sanders, M. (2007) Utility of imaging mass spectrometry (IMS) by matrixassisted laser desorption ionization (MALDI) on an ion trap mass spectrometer in the analysis of drugs and metabolites in biological tissues. J Pharm Toxicol Meth, 55, 279–288. 4. Chaurand, P., Schwartz, S. A., Billheimer, D., Xu, B. J., Crecelius, A., Caprioli, R. M. (2004) Integrating histology and imaging mass spectrometry. Anal Chem, 76, 1145–1155. 5. Maddalo, G., Petrucci, F., Iezzi, M., Pannellini, T., Del Boccio, P., Ciavardelli, D., Biroccio, A., Forli, F., Di Ilio, C., Ballone, E., Urbani, A., Federici, G. (2005) Analytical assessment of MALDI-TOF imaging mass spectrometry on this histological samples. An insight into proteome investigation. Clin Chim Acta, 357, 210–218. 6. Garrett, T. J., Prieto-Conaway, M. C., Kovtoun, V., Bui, H., Izgarian, N., Stafford, G. C., Yost, R. A. (2007) Imaging of small molecules in tissue sections with a new intermediate-pressure MALDI linear ion trap mass spectrometer. Int J Mass Spectrom, 260, 166–176. 7. Garrett, T. J., Yost, R. A. (2006) Analysis of intact tissue by intermediate-pressure MALDI on a linear ion trap mass spectrometer. Anal Chem, 78, 2465–2469. 8. Reyzer, M. L., Hsieh, Y., Ng, K., Korfmacher, W. A., Caprioli, R. M. (2003) Direct analysis of drug candidates in tissue by matrix-assisted laser desorption/ionization mass spectrometry. J Mass Spectrom, 38, 1081–1092.

9. Jackson, S. N., Wang, H.-Y. J., Woods, A. S. (2005) In situ structural characterization of phosphatidylcholines in brain tissue using MALDI-MS/MS. J Am Soc Mass Spectrom, 16, 2052–2056. 10. Winograd, N. (2003) Prospects for imaging TOF-SIMS: from fundamentals to biotechnology. Appl Surface Sci, 203–204, 13–19. 11. Caprioli, R. M., Farmer, T. B., Gile, J. (1997) Molecular imaging of biological samples: localization of peptides and proteins using MALDI-TOF MS. Anal Chem, 69, 4751–4760. 12. Todd, P. J., Schaaf, T. G., Chaurand, P., Caprioli, R. M. (2001) Organic ion imaging of biological tissue with secondary ion mass spectrometry and matrix-assisted laser desorption/ionization. J Mass Spectrom, 36, 355–369. 13. Wang, H.-Y. J., Jackson, S. N., McEuen, J., Woods, A. S. (2005) Localization and analysis of small drug molecules in rat brain tissue sections. Anal Chem, 77, 6682–6686. 14. Troendle, F. J., Reddick, C. D., Yost, R. A. (1999) Detection of pharmaceutical compounds in tissue by matrix-assisted laser desorption/ionization and laser desorption/chemical ionization tandem mass spectrometry with a quadrupole ion trap. J Am Soc Mass Spectrom, 10, 1315–1321. 15. Jackson, S. N., Wang, H.-Y. J., Woods, A. S. (2005) Direct profiling of lipid distribution in brain tissue using MALDI-TOFMS. Anal Chem, 77, 4523–4527. 16. Sjovall, P., Lausmaa, J., Johansson, B. (2004) Mass spectrometric imaging of lipids in brain tissue. Anal Chem, 76, 4271–4278. 17. Touboul, D., Kollmer, F., Niehuis, E., Laprevote, O., Brunelle, A. (2005) Improvement of biological time-of-flight-secondary ion mass spectrometry imaging with a bismuth cluster ion source. J Am Soc Mass Spectrom, 16, 1608–1618. 18. Cha, S., Yeung, E. S. (2007) Colloidal graphite-assisted laser desorption/ionization mass spectrometry and MSn of small molecules. 1. Imaging of cerebrosides directly from rat brain tissue. Anal Chem, 79, 2373–2385. 19. Todd, P. J., McMahon, J. M., McCandlish, C. A. (2004) Secondary ion images of the

230

20.

21.

22.

23.

24.

25.

Garrett and Yost developing rat brain. J Am Soc Mass Spectrom, 15, 1116–1122. Forlenza, O. V., Schaeffer, E. L., Gattaz, W. F. (2007) The role of phospholipase A(2) in neuronal homeostasis and memory formation: implications for the pathogenesis of Alzheimer’s disease. J Neural Trans, 114, 231–238. Pettegrew, J., Panchalingam, K., Hamilton, R., McClure, R. (2001) Brain membrane phospholipid alterations in Alzheimer’s disease. Neurochem Res, 26, 771–782. Ross, B. M., Mamalias, N., Moszczynska, A., Rajput, A. H., Kish, S. J. (2001) Elevated activity of phospholipid biosynthetic enzymes in substantia nigra of patients with Parkinson’s disease. Neuroscience, 102, 899–904. Martinez, M., Mougan, I. (1998) Fatty acid composition of human brain phospholipids during normal development. J Neurochem, 71, 2528–2533. Xiao, Y., Huang, Y., Chen, Z.-Y. (2005) Distribution, depletion and recovery of docosahexanoic acid are region-specific in rat brain. Brit J Nutr, 94, 544–550. Han, X., Gross, R. W. (1995) Structural determination of picomole amounts of phospholipids via electrospray ionization tandem mass spectrometry. J Am Soc Mass Spectrom, 6, 1201–1210.

26. Garrett, T. J. (2006) Ph.D. Dissertation, University of Florida, Department of Chemistry, Gainesville, FL. 27. Hsu, F.-F., Bohrer, A., Turk, J. (1997) Formation of lithiated adducts of glycerophosphocholine lipids facilitates their identification by electrospray ionization tandem mass spectrometry. J Am Soc Mass Spectrom, 9, 516–526. 28. Hsu, F.-F., Turk, J. (2000) Structural determination of sphingomyelin by tandem mass spectrometry with electrospray ionization. J Am Soc Mass Spectrom, 11, 437–449. 29. Agranoff, B. W., Benjamins, J. A., Hajra, A. A. (1998) in Basic neurochemistry. Molecular, cellular and medical aspects (Siegel, G. J., Agranoff, B. W., Fisher, S. K., Albers, R. W., Uhler, M. D., Eds.), 47–67, Lippincott– Raven, Philadelphia, PA. 30. Hsu, F.-F., Turk, J. (2005) Studies on phosphatidylserine by tandem quadrupole and multiple stage quadrupole ion-trap mass spectrometry with electrospray ionization: structural characterization and the fragmentation processes. J Am Soc Mass Spectrom, 16, 1510–1522. 31. McClellan, J. E., Murphy, I., James P., Mulholland, J. J., Yost, R. A. (2002) Effects of fragile ions on mass resolution and on isolation for tandem mass spectrometry in the quadrupole ion trap mass spectrometer. Anal Chem, 74, 402–412.

Chapter 13 Chemical Imaging with Desorption Electrospray Ionization Mass Spectrometry Vilmos Kertesz and Gary J. Van Berkel Abstract Desorption electrospray ionization mass spectrometry (DESI-MS) is a relatively new ambient surface analysis method. This technique requires little or no sample preparation prior to the actual analysis. Along with other application areas, DESI-MS is used to obtain chemical images of different biomolecules in biological tissue sections. In the case of biological tissue analysis, only preparation of the tissue slices is necessary. Furthermore, the method does not require the use of an ionization matrix preventing the redistribution of analytes prior to the analysis. Chemical images are obtained by monitoring the mass spectrometric signals of compounds while moving the sample relative to the DESI sprayer. Key words: Mass spectrometry imaging, desorption electrospray ionization, whole body, tissue section, propranolol.

1. Introduction The first publication on desorption electrospray ionization mass spectrometry (DESI-MS) appeared in 2004 (1). That paper has been followed by a stream of publications on a variety of atmospheric pressure surface sampling/ionization techniques for mass spectrometry (2). The popularity of DESI can be owed to its ease of use. In DESI, charged solvent droplets from a pneumatically assisted electrospray ionization source impact the surface to be analyzed, desorbing and ionizing analytes. The gas-phase analyte ions are then collected and transferred into the inlet of the mass spectrometer either directly or by using a transfer tube.

S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_13, © Springer Science+Business Media, LLC 2010

231

232

Kertesz and Van Berkel

A potentially very useful application for DESI-MS is chemical imaging. DESI-MS has been used to image inked lettering and shapes on paper (3–6), dyes (7), and tryptic peptide digests (8) separated on TLC plates; latent fingerprints on glass (9), endogeneous lipids (3, 10, 11), and dosed clozapine and metabolites (11, 12) in rat organ thin tissue sections; and dosed propranolol in mouse whole-body thin tissue sections (13). In this chapter we describe the method to obtain DESI-MS chemical images of whole-body thin tissue sections.

2. Materials 2.1. Chemicals

1. HPLC-grade water. 2. HPLC-grade methanol. 3. Sodium chloride (anal. grade, Sigma-Aldrich, St. Louis, MO, USA). 4.

D , L -Propranolol

hydrochloride from Sigma-Aldrich.

5. Liquid nitrogen. 6. Solid carbon dioxide (dry ice). 7. Isoflurane (Abbott Laboratories, Abbot Park, IL, USA). 8. Hexane from Sigma-Aldrich. 2.2. Biological Tissues

1. Male CD-1 mouse (Charles River Laboratories, Shrewsbury, MA, USA). 2. Whole-body thin tissue section (see Section 3.1). 3. 2% aqueous carboxymethyl cellulose embedding media (Sigma-Aldrich, part no.: C4888, carboxymethyl cellulose sodium salt, medium viscosity, 400–800 cP). 4. 3×4 in., 1.2 mm thick glass slide (Brain Research Laboratories). 5. Photoactivated polymer adhesive (Macro-Tape-Transfer System, Instrumedics, St. Louis, MO, USA).

2.3. Equipment

1. Cryomicrotome (Model CM3600, Leica Microsystems, Inc., Bannockburn, IL, USA). 2. Vacuum desiccator (Bioworld, Dublin, OH, USA). 3. Flat-bed scanner (HP Scanjet 4370 Hewlett-Packard, Palo Alto, CA, USA). 4. Syringes (Hamilton, Reno, NV, USA). 5. Syringe pump (Model 22, Harvard Apparatus, Cambridge, MA, USA).

Chemical Imaging with DESI-MS

233

6. Stainless steel union (Part no. U-412, Upchurch Scientific, Inc., Oak Harbor, WA, USA). 7. Teflon tubing (approximately 40 cm of 100 µm i.d., 1/16 in. o.d., part no. 1474 from Upchurch Scientific, Inc.). 8. 5.2 cm long taper-tip fused-silica capillary (50 µm i.d., 360 µm o.d.; from New Objective, Woburn, MA, USA). 9. Microionspray II head (MDS Sciex, Concord, ON, Canada). 10. Particle discriminator interface (PDI) from MDS Sciex (14). 11. Mass spectrometer (4000 QTRAP from MDS Sciex). 12. 360◦ manual rotational stage (Newport, Irvine, CA, USA). 13. XYZ robotic platform (capable of 100 mm × 100 mm travel in the XY plane, Model MS2000, Applied Scientific Instrumentation, Inc., Eugene, OR, USA) (see Note 1). 14. Computer-controlled communication device with at least one digital output (DO) port enabled to send a TTL signal to trigger MS data acquisition (Model USB-1208LS USBbased data acquisition module, Measurement Computing Corp., Norton, MA, USA). 15. CCD camera and associated monitor (Protana A/S, Odense, Denmark).

3. Methods 3.1. Tissue Preparation

1. Administer propranolol intravenously to a mouse via the tail vein at 7.5 mg/kg as an aqueous solution in 0.9% NaCl (2 ml/kg of a 3.75 mg/ml propranolol solution). Euthanize the mouse 20 or 60 min after the infusion with an isoflurane overdose and immediately freeze in dry ice/hexane. Embed the frozen mouse in 2% aqueous carboxymethyl cellulose. 2. Prepare sagittal whole-body cryosections (40 µm thick) using a cryomicrotome. Transfer the frozen sections to glass slides using a tape-transfer process, which utilizes a photoactivated polymer adhesive. After transfer to the glass slides, freeze-dry the sections within the chamber of the cryomicrotome. Store all tissue sections in a desiccator at room temperature until analysis. 3. Color images of the tissue sections were acquired using a flat-bed scanner.

234

Kertesz and Van Berkel

3.2. DESI-MS Setup

1. The DESI emitter used for these experiments was identical to that used in (7). The spray emitter was a 5.2 cm long, taper-tip, fused-silica capillary with 50 µm i.d. and 360 µm o.d. in a microionspray head with a 500 µm i.d. nebulizing gas tube providing a nebulizing gas (nitrogen) jet annulus area of approximately 1.5 × 10−7 m2 (see Note 2). 2. Mount the DESI emitter on a 360◦ manual rotational stage to allow adjustment of the incident angle. 3. Set up the XYZ robotic platform. The travel capacity of the platform should allow analysis of the whole tissue section (see Note 3). 4. Mount the PDI with 7 cm heated chamber length (sampling capillary) on the mass spectrometer (see Note 4). 5. Mount the CCD camera and install the monitor to observe the DESI emitter and sampling capillary from an angle just above the surface (see Note 5). 6. While the instrument is in standby or off (high voltage is not applied), couple the high-voltage cable to the stainless steel body of the microionspray head and place a grounded stainless steel union in the transfer line (Teflon tubing) between the syringe pump and the DESI emitter (see Note 6). 7. Optimize mass spectrometer settings and selected reaction monitoring (SRM) transitions of the compounds you wish to monitor during chemical imaging by continuous infusion of approximately 0.1 µM (depending on the ionization efficiency) test solutions in an on-axis position about 1 cm from the entrance of the PDI. Optimized emitter voltage in our setup was 4 kV with a nebulizer gas setting of 50 (linear velocity of 282 m/s), while monitored SRM transitions were m/z 260 → 116 and m/z 260 → 183 with a collision energy of 27 eV for propranolol. 8. Optimize DESI setup geometry used during DESI imaging: spray a 0.1 µM test solution of a compound you wish to monitor onto a glass surface and monitor the SRM signal(s) of it while changing the spray tip-to-surface and spray tip-toPDI entrance distances and the incident angle. Optimized setup for propranolol in our laboratory included positioning the emitter from approximately 2 mm from both the surface to be analyzed and from the entrance of the PDI, at an approximately 55◦ angle to the surface. 9. Wire a DO port of the computer-controlled communication device to the appropriate digital input (DI) port of the AUX control interface of the mass spectrometer. Caution: Inappropriate wiring can permanently damage the AUX interface!

Chemical Imaging with DESI-MS

3.3. DESI-MS Imaging

235

During an imaging experiment, the surface was moved as follows. At the beginning of an imaging experiment the spray impact plume was positioned at one corner of the area of interest and the sampling end of the PDI-heated chamber was positioned to just touch the tissue. The first lane was scanned by moving the surface parallel to the x-axis at a forward surface scan rate with the spray plume fixed in the optimal position in front of the sampling inlet. At the end of the first lane, the surface was lowered by 2 mm followed by moving the surface at a return speed back to the beginning of the first lane. With the surface lowered there was a 2 mm gap between the sampling capillary and the surface. As such, the desorption spray plume impacted the tissue surface outside the area of interest or at an already analyzed area preventing contamination of yet unanalyzed tissue regions. At the same time, the 2 mm gap between the sampling capillary and the surface eliminated ion signal as well. When the beginning of the first lane was reached, the surface was moved parallel to the y-axis with the lane spacing distance followed by raising the surface 2 mm so the sampling capillary touched the tissue again. The following lanes were scanned similar to the first lane. 1. Mount the glass slide holding the tissue section onto the robotic platform using a double-sided tape (see Note 7). Be sure that the sample is secured and completely parallel to the XY plane of the robotic platform. 2. Decide on the surface scan rate, the size of the area to be imaged, and the lane spacing. Set up these parameters in the stage control software and instruct the software to send a trigger signal to the mass spectrometer at the beginning of each lane scan. 3. Calculate the time necessary to scan a lane using the following formula: acquisition time per lane (ATPL) = lane length/surface scan rate. Calculate the time (TB) necessary to move the surface to the starting position of the following lane after a lane is scanned. This time is the sum of the time necessary to lower the surface by 2 mm at the end of the lane scanned, to move the surface at a return speed back to the beginning of the lane scanned, to move the surface parallel to the y-axis with the lane spacing distance, and to raise the surface 2 mm so the sampling capillary touches the tissue again. The acquisition time of the whole experiment can be calculated as number of lane times (ATPL + TB). 4. Fill up a glass syringe with the appropriate amount of DESI solvent. For propranolol-dosed tissues 80/20 (v/v) methanol/water solution was delivered to the emitter by a syringe pump at a flow rate of 5 µl/min (see Note 8). Make sure that the syringe contains enough DESI solvent to

236

Kertesz and Van Berkel

acquire the chemical image of the desired size (see Note 9). This can be calculated by multiplying the flow rate and the acquisition time of the whole experiment and adding some extra solvent to allow for initial spray stabilization. 5. Create a new sample list in the mass spectrometer acquisition software. Set the number of samples in the list equal to the number of lanes to be scanned. Set the acquisition time of a sample to be equal to ATPL. Instruct the software to wait for an external trigger signal to start acquiring data before each sample. 6. Move the surface so when spraying on it the analyzed areas of interest are not contaminated (see Note 10). Start the syringe pump, turn on the nebulizer gas and high voltage, and let the spray stabilize for approximately 1 min before starting. 7. Move the surface so the spray plume interrogates the starting point. Make sure that the PDI just touches the surface (see Fig. 13.1 and Note 11).

DESI emitter

surface

PDI

Fig. 13.1. Photograph of a 4000 QTRAP mass spectrometer atmospheric sampling interface region equipped with a PDI during imaging of a mouse thin tissue section showing the DESI emitter, surface, and sampling inlet. (Reproduced from ref. (13) with permission from American Chemical Society.)

8. Start data acquisition in the mass spectrometer software. The mass spectrometer should wait for a trigger signal from the stage control software. 9. Start the movement of the surface in the stage control software. 10. When all sample files are collected (see Note 12), turn off the syringe pump, wait for approximately 1 min (to stop the solvent delivery), and then turn off the nebulizer gas and the high voltage. Lower the surface so the PDI is above

Chemical Imaging with DESI-MS

237

the surface by approximately 2 mm. This procedure ensures that even if the spray is turned on accidentally or removal of the slide is attempted, it will not contaminate an area of interest or damage the spray tip, respectively. 11. As a first step of the data analysis, convert the data files into a format that is appropriate to the visualization software of your choice. For example, the freely available ImgConverter v3.0 software (11) is able to convert Xcalibur mass spectra files (.RAW extensions) into Analyze 7.5 format files (.img, .hdr, and .t2m) that are required by the BioMap image analysis software (http://www.maldimsi.org, free of charge). BioMap is then able to generate 2D images of the surface versus the intensity. For Sciex platforms, we have developed an analyst processing script that converts analyst data files (.WIFF extensions) with SRM transitions into a tab-separated text file that can be easily imported into any popular data analysis and graphing software [e.g., Origin (www.originlab.com) and SigmaPlot (www.sigmaplot.com)] to plot as a 2D surface plot (intensity versus X, Y coordinates) (see Note 13). Figures 13.2 and 13.3 show distribution of propranolol in mouse thin tissue sections obtained using surface scan rates of 0.1 and 7 mm/s, respectively.

(a) Brain

Lung

Liver

Kidney

Stomach contents

10 mm (b)

0

100

Fig. 13.2. (a) Scanned optical image of a 40 µm thick sagittal whole-body tissue section of a mouse dosed intravenously with 7.5 mg/kg propranolol and euthanized 20 min after dose. (b) Distribution of propranolol in 20×20 mm and 38×20 mm areas measured by DESI-MS/MS (SRM: m/z 260 → 116) using 80/20 (v/v) methanol/water as DESI solvent at a flow rate of 5 µl/min. Surface scan rate was 0.1 mm/s, dwell time was 100 ms, and the images were created from 41 lanes with 500 µm spacing. Pixel size was 84 µm (h) × 500 µm (v), and experiment times were 150 and 285 min for the 20×20 mm and 38×20 mm areas, respectively. (Reproduced from ref. (13) with permission from American Chemical Society.)

238

Kertesz and Van Berkel

Kidney

Lung

Brain

Heart

(a)

Liver Stomach contents

10 mm

(b)

0

100

Fig. 13.3. (a) Scanned optical image of a 40 µm thick sagittal whole-body tissue section of a mouse dosed intravenously with 7.5 mg/kg propranolol and euthanized 60 min after dose. (b) Distribution of propranolol in the 94 mm × 30 mm tissue section presented in (a) measured by DESI-MS/MS (SRM: m/z 260 → 116) using 80/20 (v/v) methanol/water as DESI solvent at a flow rate of 5 µl/min. Surface scan rate was 7 mm/s, dwell time was 20 ms, and the image was created from 151 lanes with 200 µm spacing. Pixel size was 140 µm (h) × 200 µm (v), and total experiment time was 79 min. (Reproduced from ref. (13) with permission from American Chemical Society.)

4. Notes 1. The controller of the MS2000 robotic platform is also equipped with a joystick that allows manual control of the stage. This feature is extremely useful for quick manual positioning, simple testing, etc. 2. Check if the tip of the fused silica capillary is not broken or cracked. A damaged tip most likely results in a skewed spray causing irreproducible data. Replace the tip if necessary. 3. While in the most common configuration the XY plane is horizontal, vertical configuration of the XY plane can be advantageous sometimes, i.e., for easier access to the sample, monitoring purposes. 4. PDIs with different lengths (e.g., 2 cm) are available. Use of the shortest PDI to analyze the whole surface minimizes ion-transfer losses.

Chemical Imaging with DESI-MS

239

5. A zoomed view of the spray tip/plume/sampling capillary region is critical when setting up the ideal DESI conditions before an experiment and when keeping the sampling capillary on the surface during the chemical imaging experiment. 6. Caution: The DESI emitter floats at the high ES voltage, and appropriate shields and interlocks should be used to avoid accidental contact with this component. 7. Analyzing tissue samples with DESI can create harmful “bioaerosols.” Appropriate safety protocols should be followed and protective equipment should be worn with respect to exposure of the operator and decontamination of equipment used in the DESI imaging studies. 8. The DESI solvent, and solvent and gas flow rates, should be optimized for the specific tissue and analyte. Under optimal conditions, only the most energetic part of the DESI spray plume should be able to extract the analyte out of the tissue, otherwise the “washing effect” (15) will result in cross contamination and signal and resolution loss. Also, if the gas and/or the solvent flow rates are high, it can extensively damage the tissue section. It is advised to sacrifice a tissue section to optimize these parameters for maximum ion signal and resolution. 9. Use of the combination of a syringe and a syringe pump that delivers a stable solvent flow is important for successful DESI imaging. Alternatively, deliver the solvent with an HPLC pump that has sufficient solvent reservoir capacity. 10. It is extremely important not to contaminate the area of interest before a DESI imaging. Be sure that there is no spray or the spray plume moves outside of the area of interest when moving the surface. 11. Using a PDI, the highest ion signal can be achieved by the PDI just touching the surface. However, with other sampling capillary configurations, e.g., using an extended ion-transfer tube available from Prosolia (Indianapolis, IN, USA) for Thermo instruments, the optimal position of the sampling capillary may be different. 12. It is advised to at least periodically check on the data in the files that are collected. For example, timely recognition of accidental clogging of the sampling capillary caused by tissue particles and indicated by DESI signal loss can save valuable research time. 13. Because of the wide range of the mass spectrometers that can be used for DESI imaging and the data formats they use, only an example of the data analysis process was given.

240

Kertesz and Van Berkel

Most mass spectrometer manufacturers distribute a selfdevelopment kit (SDK) to access the data in the data files they support (e.g., Xcalibur SDK from Thermo, Analyst Scripting from Sciex). See the mass spectrometer manufacturer about these SDKs.

Acknowledgments The imaging protocols described here were derived from work supported by the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences, United States Department of Energy under Contract DE-AC05-00OR22725 with ORNL, managed and operated by UT-Battelle, LLC. References 1. Takáts, Z., Wiseman, J. M., Gologan, B., Cooks, R. G. (2004) Mass spectrometry sampling under ambient conditions with desorption electrospray ionization. Science, 306, 471–473. 2. Van Berkel, G. J., Pasilis S. P., Ovchinnikova, O. (2008) Established and emerging atmospheric pressure surface sampling/ionization techniques for mass spectrometry. J Mass Spectrom, 43, 1161–1180. 3. Ifa, D. R., Wiseman, J. M., Song, Q., Cooks, R. G. (2007) Development of capabilities for imaging mass spectrometry under ambient conditions with desorption electrospray ionization (DESI). Int J Mass Spectrom, 259, 8–15. 4. Kertesz, V., Van Berkel, G. J. (2008) Scanning and surface alignment considerations in chemical imaging with desorption electrospray mass spectrometry. Anal Chem, 80, 1027–1032. 5. Ifa, D. R., Gumaelius, L. M., Eberlin, L. S., Manicke, N. E., Cooks, R. G. (2007) Forensic analysis of inks by imaging desorption electrospray ionization (DESI) mass spectrometry. Analyst, 132, 461–467. 6. Kertesz, V., Van Berkel, G. J. (2008) Improved imaging resolution in desorption electrospray ionization mass spectrometry. Rapid Commun Mass Spectrom, 22, 2639–2644. 7. Van Berkel, G. J., Kertesz, V. (2006) Automated sampling and imaging of ana-

8.

9.

10.

11.

12.

13.

lytes separated on thin-layer chromatography plates using desorption electrospray ionization mass spectrometry. Anal Chem, 78, 4938–4944. Pasilis, S. P., Kertesz, V., Van Berkel, G. J., Schulz, M., Schorcht, S. (2008) HPTLC/DESI-MS imaging of tryptic protein digests separated in two dimensions. J Mass Spectrom, 43, 1627–1635. Ifa, D. R., Manicke, N. E., Dill, A. L., Cooks, R.G. (2008) Latent fingerprint chemical imaging by mass spectrometry. Science, 321, 805. Wiseman, J. M., Ifa, D. R., Song Q., Cooks, R. G. (2006) Tissue imaging at atmospheric pressure using desorption electrospray ionization (DESI) mass spectrometry. Angew Chem Int Ed, 45, 7188–7192. Wiseman, J. M., Ifa, D. R. Venter, A., Cooks, R. G. (2008) Ambient molecular imaging by desorption electrospray ionization mass spectrometry, Nat Protoc, 3, 517–524. Wiseman, J. M., Ifa, D. R., Zhu, Y., Kissinger, C. B., Manicke, N. E., Kissinger, P. T., Cooks, R. G. (2008) Desorption electrospray ionization mass spectrometry: imaging drugs and metabolites in tissues. Proc Nat Acad Sci U S A, 105, 18120–18125. Kertesz, V., Van Berkel, G. J., Vavrek, M., Koeplinger, K. A., Schneider, B. B., Covey, T. R. (2008) Comparison of drug distribution images from whole-body thin tissue sections

Chemical Imaging with DESI-MS obtained using desorption electrospray ionization tandem mass spectrometry and autoradiography. Anal Chem, 50, 5168–5177. 14. Leuthold, L. A., Mandscheff, J.-F., Fathi, M., Giroud, C., Augsburger, M., Varesio, E., Hopfgartner, G. (2006) Desorption electrospray ionization mass spectrometry: direct

241

toxicological screening and analysis of illicit ecstasy tablets. Rapid Commun Mass Spectrom, 20, 103–110. 15. Pasilis, S. P., Kertesz, V., Van Berkel, G. J. (2008) Surface scanning analysis of planar arrays of analytes with desorption electrospray ionization-mass spectrometry. Anal Chem, 79, 5956–5962.

Chapter 14 Mass Spectrometry Imaging of Small Molecules Using Matrix-Enhanced Surface-Assisted Laser Desorption/Ionization Mass Spectrometry (ME-SALDI-MS) Qiang Liu, Yongsheng Xiao, and Lin He Abstract Surface-assisted laser desorption/ionization mass spectrometry (SALDI-MS) uses inorganic particles or porous surfaces as the energy-mediating means to promote desorption and ionization of low-mass analytes of interest. With good stability during laser ablation, SALDI substrates exhibit reduced background in the low-mass region that is often crowded in conventional matrix-assisted laser desorption/ionization (MALDI) due to matrix fragmentation; a benefit renders SALDI-MS attractive in imaging low-mass species. Practical application of SALDI-MS, however, is hindered by its unsatisfied detection sensitivity for most compounds. With aims of improving MS imaging resolution and sensitivity of low-mass species, we describe an experimental protocol using a hybrid ionization method, termed matrix-enhanced SALDI (ME-SALDI), to detect crucial low-mass species with their spatial distribution in mouse brain tissue. Key words: Surface-assisted laser desorption/ionization (SALDI), desorption ionization on porous Si (DIOS), matrix-enhanced SALDI (ME-SALDI), low-mass species, metabolite, 2D imaging.

1. Introduction Surface-assisted laser desorption/ionization is a laser-induced ionization method developed in parallel to the more well-known matrix-assisted laser desorption/ionization (MALDI) method. Using inorganic materials as energy-mediating media, it has been exploited as a possible alternative to MALDI, especially in small molecule detection (1–4). The most successful SALDI substrate was reported by Siuzdak and coworkers, in which a silicon surface of mesoporous features was used, commonly termed as S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_14, © Springer Science+Business Media, LLC 2010

243

244

Liu, Xiao, and He

desorption/ionization on porous silicon (DIOS) (4). The use of DIOS-MS in 2D MS imaging of metabolites has been successfully demonstrated (5). Other SALDI-MS imaging of metabolites using different SALDI substrates has also been reported in the literature (6, 7). With the aim of improving MS imaging sensitivity, SALDI and MALDI are synergistically combined to yield a hybrid ionization method, i.e., matrix-enhanced surface-assisted laser desorption ionization (ME-SALDI) (8). Comparing to conventional MALDI, this ionization source uses the porous Si surface as an effective photon absorption medium that significantly reduces laser fluence needed to bring small molecules into the gas phase; consequently less fragmentation is observed. The porous surface playing the major role in desorption also renders the desorption process less dependent on the “perfect” incorporation of matrix and analyte molecules. As a result, a much smaller amount of matrix can be applied to achieve better detection sensitivity with reduced background. In comparison to SALDI/DIOS, addition of acidic matrix molecules provides a proton-rich environment that drastically improves the ionization efficiency of the desorbed species (i.e., better sensitivity) (9) and expands the detectable mass range. Localized extraction of analytes due to the employment of matrix solutions of traditional MALDI is upheld as the extra bonus to traditional SALDI.

2. Materials 2.1. Preparation of Porous SALDI/ DIOS Substrates

1. n-Type Sb-doped (100) single-crystalline silicon wafers at 0.005–0.02 /cm (Silicon Sense Inc., Nashua, NH, USA) were stored under vacuum prior to use. 2. Hydrofluoric acid (HF) etching solution: 49% HF (Fisher Scientific) with 95% ethanol (Aaper Alcohol) (v/v 1:1). 3. Hydrofluoric acid (HF) wash solution: 49% HF (Fisher Scientific) with 95% ethanol (Aaper Alcohol) (v/v 1:10). 4. Hydrogen peroxide (H2 O2 ) oxidation solution: 30% H2 O2 (Fisher Scientific) with 95% ethanol (Aaper Alcohol) (v/v 1:1).

2.2. Tissue Sections Preparation

1. Optimal Cutting Temperature (OCT) compound (Ted Pella Inc., Redding, CA, USA). 2. Methylene blue (MP Biomedicals) staining solution: 0.15 g methylene blue powder in 100 ml of 70% ethanol. The solution was stirred overnight.

Mass Spectrometry Imaging of Small Molecules

245

3. Leica Stainless-Steel Re-usable Microtome Knives (Fisher Scientific) were used in tissue slicing. 4. Carbon steel surgical blades (Bard-Parker) were used to precut tissue samples into small portions. 5. Mouse brain tissue was received as gift. 6. Cryo-cut microtome (American Optical Corp., Buffalo, NY, USA). 2.3. Matrix Deposition

1. 2,5-Dihydroxybenzoic acid (DHB)and α-cyano-4hydroxycinnamic acid (CHCA) (Sigma-Aldrich, St. Louis, MO, USA). 2. Silicon oil (Sigma-Aldrich) was used in the oil bath. 3. Sublimation glassware (Chemglass). 4. Edwards E2M8 high vacuum pump (Edwards High Vacuum, Burlington, ON, Canada) with a vacuum meter. 5. EG&G Princeton Potentiostat, Model 273 (EG&G Princeton Applied Research, Princeton, NJ, USA).

2.4. Mass Spectrometry

1. Applied Biosystems Voyager DE-STR MALDI-TOF mass spectrometer (Framingham, MA, USA). 2. MMSIT MALDI Imaging Tool software V2.2.0 (© 2004 by Dr. Markus Stoeckli, Novartis Institutes for BioMedical Research, Basel, Switzerland) downloadable at www.maldimsi.org. 3. BioMap 3.7.5.4 (by Dr. Markus Stoeckli, Novartis Institutes for BioMedical Research).

2.5. Optical Microscope

1. Leica DMRX light microscope (Leica Microsystems, Wetzlar, Germany) equipped with a Donpisha XC-003P CCD (Sony, Japan).

3. Methods 3.1. Preparation of Porous SALDI Substrate

1. Silicon wafers were cut into 1-cm2 square-shaped chips and dipped into the HF wash solution for 1 min to remove the oxidized layer. (Caution: HF can cause severe tissue damage upon contact or inhalation.) The Si chips were then washed in 95% ethanol and dried with N2 . 2. A Si chip was assembled in an anodic etching cell made of Teflon, as shown in Fig. 14.1. A three-electrode system was used where a Au working electrode was placed under the Si chip, a Pt counter electrode, and a Pt reference electrode

246

Liu, Xiao, and He

Fig. 14.1. A schematic drawing of an electrochemical etching cell for porous Si preparation.

were placed above the surface. The cell was clamped tight before it was filled with the HF etching solution. A 50-W tungsten lamp was aligned to provide uniform illumination of the Si surface. An EG&G Princeton Potentiostat, Model 273, was programmed to control etching current and time. The complete setup was placed in a chemical hood for ventilation (see Note 1). 3. The chip was electrochemically etched in the HF etching solution for 1 min at a current density of 5 mA/cm2 . HF waste was carefully transferred to a pre-labeled waste bottle in the hood dedicated to HF. The setup was washed with 95% ethanol before disassembly. The produced porous Si substrate (i.e., DIOS) was washed again with copious 95% ethanol, dried with N2 . 4. Double-etch of the DIOS substrate was carried out by first dipping the substrate into the H2 O2 oxidation solution for 1 min, followed by washing in 95% ethanol, drying in a N2 stream, and dipping in the HF wash solution. The substrate was then washed again in 95% ethanol and stored in the same medium till needed (see Note 2). 5. Prior to usage, the substrates were refreshed in the HF wash solution for 1 min, washed in 95% ethanol, and dried with N2 .

Mass Spectrometry Imaging of Small Molecules

3.2. Tissue Section Preparation

247

1. Fresh mouse heart and liver tissues were pre-cut into thin slices using carbon steel surgical blades with at least one dimension at less than 2 mm. Fresh mouse brain was used directly without pre-cutting. All tissue or thin tissue slices were snap frozen in liquid N2 immediately and stored in –80◦ C prior to usage (see Note 3). 2. A cryo-cut microtome equipped with Leica Stainless-Steel Re-usable Microtome Knives was used in the following description. Procedural modification may be needed when different equipment was to be used. Frozen animal tissue slices were mounted onto the microtome chunk that held the tissue in place with the OCT compound. The temperature was slowly brought up to −20◦ C. The frozen tissue samples were then sectioned into 10-µm-thick slices. The 10-µm-thick tissue slices were transferred onto a dry porous Si substrate and slowly brought to room temperature. 3. Meanwhile, the adjacent section was transferred onto a glass slide for conventional histological staining in which the methylene blue staining solution (0.15%, 200 µl) was deposited onto the tissue section. After 10-s staining, excess stain was washed away with 95% ethanol. 4. A Leica DMRX light microscope equipped with a Donpisha XC-003P CCD camera was used to collect the optical image of stained sections to provide visual validation of IMS data.

3.3. Matrix Deposition

1. A sublimation apparatus was set up in a chemical hood (Fig. 14.2). An Edwards E2M8 high vacuum pump with a vacuum meter to provide controllable vacuum in the sublimation chamber (see Note 4). 2. The tissue coated porous Si was attached to the bottom of the apparatus condenser in direct contact with the running cooling water. 0.3 g of CHCA or DHB was added to the bottom of the sublimation chamber (see Note 5). 3. The oil bath was pre-heated to 110ºC for DHB or to 130ºC for CHCA. The chamber pressure was maintained at ∼50 torr for 2 min before immersing the apparatus into oil bath for 5 min. 4. After removing the sublimation setup from the oil bath, the vacuum was released and the DIOS coated with matrix was taken out and loaded into the MS sample chamber immediately. The vacuum pump, the cooling water, and the oil bath were turned off. Excess DHB and CHCA were discarded.

248

Liu, Xiao, and He

Cooling H2O out Cooling H2O in

Vacuum

SALDI substrate with Matrix

tissue atop

Fig. 14.2. A photo picture of in-house sublimation apparatus.

3.4. Mass Spectrometry Imaging

1. A Voyager DE-STR MALDI-TOF mass spectrometer was used for the following description. Modification in instrumental conditions may be needed to achieve optimal imaging results when different instruments are used. 2. The mass spectrometer was equipped with a 20-Hz N2 laser. The laser irradiation energy was adjusted by a neutral density filter and the beam size was adjusted to 35 µm with an adjustable pinhole placed close to the laser entrance window. The actual laser beam size was obtained by increasing laser fluence till a burn mark was left behind to allow off-line measurements of the laser beam footprint (see Note 6). 3. The matrix-coated DIOS substrate with a piece of mouse brain tissue attached was mounted on a stainless steel MALDI plate with double-side conductive tape. 4. The instrument was operated at an accelerating voltage of 20 kV in a reflector and in a positive-ion mode. 50 Laser shots were averaged at each location to yield one accumulated spectrum for each imaging pixel. The translational stage was operated at 50-µm stepwise (see Note 7).

Mass Spectrometry Imaging of Small Molecules

249

5. The MS instrument was controlled by MMSIT MALDI Imaging Tool software. The imaging area was manually selected along the outline of the tissue section in MMSIT (see Note 8). 6. 2D ion maps were reconstructed using BioMap 3.7.5.4. An example of results is shown in Fig. 14.3.

Fig. 14.3. Optical images of the coronal sections of (a) mouse cerebrum and (g) mouse cerebellum. Representative MS spectra collected from the adjacent coronal sectioning of (b) mouse cerebrum and (h) mouse cerebellum in ME-SALDI, respectively. Also shown are reconstructed 2D images for ions at (c) m/z=369.4, (d) m/z =772.6, (e) m/z = 838.6, and (f) m/z = 844.5 from the mouse cerebrum coronal section, and for ions at (i) m/z =369.4, (j) m/z =769.6, (k) m/z = 826.6, and (l) m/z = 838.6 from mouse cerebellum coronal section. Molecular identification sees the text. Different areas in the brain tissue are labeled numerically: 1, cerebral cortex; 2, corpus callosum; 3, striatum; 4, cerebellar nuclei; 5, molecular layer in cerebellum; 6, brain stem. Figure reproduced from (8) with permission.

250

Liu, Xiao, and He

4. Notes 1. During the etching process, gas bubbles were mildly, but continuously released. The Si chip turned from gray to bright blue first, then quickly changed to a golden color. Over time the color of the Si chip turned darker and darker, suggesting formation of a rough surface. The final DIOS substrate should exhibit a dark blue hue after dried. Formation of porous features on Si was critically related to the amount of irradiating light, the current density applied, and the enduring etching time. The color of the substrate can be used as a reference point in troubleshooting. For example, a substrate with a yellowish color suggested it was over-etched whereas a light gray hue suggested it was under-etched. 2. Severe MS performance degradation has been correlated to exposure of porous Si to air. The DIOS substrates are therefore required to be stored in ethanol until needed. For the substrates coated with tissue sample, matrix deposition should occur immediately out of the same concern. 3. Freeze and thaw of tissue samples is a critical step in preparation of tissue sections. The most common problem observed in frozen tissue sections is the ice crystal damage, which causes leaky tissues and blotched tissue surface (Fig. 14.4).

Fig. 14.4. A representative optical image of a piece of mouse liver tissue section with noticeable ice crystal damages (arrows).

Mass Spectrometry Imaging of Small Molecules

251

The size of ice crystals is usually determined by the speed of the freezing process through the whole tissue. Therefore, a pre-cutting step is recommended to reduce the overall tissue size before snap freezing. Tolerance toward ice crystal damage was found at different levels for different types of tissue. For instance, the most serious ice crystal damage was observed in mouse liver sections but little in mouse brain sections. 4. Sublimation is sensitive to the vacuum pressure and the temperature (10). The conditions described here were optimized for our setup and the deposition efficiency may vary among laboratories. 5. To quantitatively control the amount and the quality of the matrix deposited, a clear glass side was coated under each sublimating condition in parallel. UV absorbance of the matrix layer on the side and the microscopic images of matrix morphology were collected to monitor the thickness and uniformity of matrix deposition. Empirically it has been found that a matrix layer with UV absorbance between 1.0 and 2.0 was, in general, suitable for ME-SALDI. 6. Irradiation energy was adjusted to achieve optimal MS performance prior to sample imaging. It was significantly lower than the energy needed for traditional MALDI. 7. Imaging resolution can be further improved through oversampling (11). 8. The overall length of an imaging experiment is mainly determined by the number of pixels (i.e., spectra, determined by the imaging spatial resolution and the tissue size), the number of laser shots per pixel, and the frequency of laser operated. For a 20-Hz laser and 50 shots/spectrum, the imaging rate is at approximately 1,440 spectra per hour, equivalent to a scanning throughput of 3.6 mm2 in an hour. The overall size of the imaging data per experiment is determined by the total number of spectra and the mass resolution and the mass window of each spectrum.

Acknowledgments We thank Dr. Dykstra at the Laboratory for Advanced Electron and Light Optical Methods at College of Veterinary Medicine, North Carolina State University (NCSU) for tissue preparation. We appreciate Drs. G. M. Pollack, J. Padowski, and W. Yue at

252

Liu, Xiao, and He

School of Pharmacy, University of North Carolina at Chapel Hill and Mrs. Welker at College of Agriculture and Life Sciences at NCSU for providing mice tissue samples. References 1. Tanaka, K., Waki, H., Ido, Y., Akita, S., Yoshida, Y., Yoshida, T. (1988) Protein and polymer analyses up to m/z 100,000 by laser ionization time-of-flight mass spectrometry. Rapid Commun Mass Spectrom, 2, 151–153. 2. Sunner, J., Dratz, E., Chen, Y.-C. (1995) Graphite surface-assisted laser desorption/ionization time-of-flight mass spectrometry of peptides and proteins from liquid solutions. Anal Chem, 67, 4335–4342. 3. Finkel, N. H., Prevo, B. G., Velev, O. D., He, L. (2005) Ordered silicon nanocavity arrays in surface-assisted desorption/ionization mass spectrometry. Anal Chem, 77, 1088–1095. 4. Wei, J., Buriak, J. M., Siuzdak, G. (1999) Desorption–ionization mass spectrometry on porous silicon, Nature, 399, 243–246. 5. Liu, Q., Guo, Z., He, L. (2007) Mass spectrometry imaging of small molecules using desorption/ionization on silicon. Anal Chem, 79, 3535–3541. 6. Taira, S., Sugiura, Y., Moritake, S., Shimma, S., Ichiyanagi, Y., Setou, M. (2008) Nanoparticle-assisted laser desorption/ionization based mass imaging

7.

8.

9.

10.

11.

with cellular resolution. Anal Chem, 80, 4761–4766. Zhang, H., Cha, S., Yeung, E. S. (2007) Colloidal graphite-assisted laser desorption/ionization ms and msn of small molecules. 2. Direct profiling and MS imaging of small metabolites from fruits. Anal Chem, 79, 6575–6584. Liu, Q., Xiao, Y., Pagan-Miranda, C., Chiu, Y. M., He, L. (2009) Metabolite imaging using matrix-enhanced surface-assisted laser desorption ionization mass spectrometry (ME-SALDI-MS). J Am Soc Mass Spectrom, 20, 80–88. Liu, Q., He, L. (2008) Semi-quantitative study of solvent and surface effects on analyte ionization in desorption ionization on silicon (DIOS) mass spectrometry. J Am Soc Mass Spectrom, 19, 8–13. Hankin, J. A., Barkley, R. M., Murphy, R. C. (2007) Sublimation as a method of matrix application for mass spectrometric imaging. J Am Soc Mass Spectrom, 18, 1646–1652. Jurchen, J. C., Rubakhin, S. S. and Sweedler, J. V. (2005) MALDI-MS imaging of features smaller than the size of the laser beam. J Am Soc Mass Spectrom, 16, 1654–1659.

Chapter 15 Preparation of Single Cells for Imaging Mass Spectrometry Elena S.F. Berman, Susan L. Fortson, and Kristen S. Kulp Abstract Characterizing the molecular contents of individual cells is critical for understanding fundamental mechanisms of biological processes. Imaging mass spectrometry (IMS) of biological systems has been steadily gaining popularity for its ability to create precise chemical images of biological samples, thereby revealing new biological insights and improving understanding of disease. In order to acquire mass spectral images from single cells that contain relevant molecular information, samples must be prepared such that cell-culture components, especially salts, are eliminated from the cell surface and that the cell contents are accessible to the mass spectrometer. We have demonstrated a cellular preparation technique for IMS that preserves the basic morphology of cultured cells, allows mass spectrometric chemical profiling of cytosol, and removes the majority of the interfering species derived from the cellular growth medium. Using this protocol, we achieve high-quality, reproducible IMS images from three diverse cell types: MCF7 human breast cancer cells, Madin-Darby canine kidney (MDCK) cells, and NIH/3T3 mouse fibroblasts. This preparation method allows rapid and routine IMS analysis of cultured cells, making possible a wide variety of experiments to further scientific understanding of molecular processes within individual cells. Key words: ToF-SIMS, imaging mass spectrometry, single cells, molecular profiling, biological imaging.

1. Introduction In the past few years, there has been an explosion in research utilizing imaging mass spectrometric techniques for biological applications (1–6). While most IMS research has focused on analysis of tissue samples, there is a clear need for single-cell analysis as well. Common cellular experiments utilize entire cell populations, thereby averaging the responses of all of the cells in the population and obscuring subtle differences among individual cells (7). S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_15, © Springer Science+Business Media, LLC 2010

253

254

Berman, Fortson, and Kulp

By interrogating single cells, an analysis is liberated from assumptions regarding cell population homogeneity and is ensured that all cellular responses can be measured. Understanding the molecular makeup of single cells is necessary to identify small cellular changes that may underlie many biological processes including disease development. There are several IMS techniques currently available for analyzing single cells. Although more commonly used for tissue analysis, recent advances in matrix-assisted laser desorption/ionization (MALDI) MS imaging have provided greatly enhanced lateral resolution, opening the potential for imaging individual cells (8, 9). To date, however, very few reports of single-cell MALDI imaging have been published (10). Secondary ion mass spectrometry (SIMS), in both dynamic and static mode, has been extensively applied to imaging analysis of single cells. Advances in imaging of cells by dynamic SIMS have been reviewed by Guerquin-Kern et al. (4), with many examples demonstrating superb spatial resolution and localization of elemental species within cells. Recently, a new generation of dynamic SIMS instrumentation, NanoSIMS, has been used to garner subcellular localization of a peptide vector (11), study diatom cell division (12), and perform nanoautography with stable isotope tracers (13) among others. Static SIMS has also been widely applied to cellular imaging, with recent examples including relative quantification of cholesterol in cell membranes (14), threedimensional imaging (15–17) (among others), and distinguishing cancerous cells of differing breast cancer phenotypes (18). Cellular imaging has also been demonstrated by a variety of other mass spectrometric techniques, including desorption/ionization on silicon (19), laser post-ionization secondary neutral mass spectrometry (20), and atmospheric pressure femtosecond laser IMS (21). We are utilizing time-of-flight secondary ion mass spectrometry (ToF-SIMS) to show the suitability of our reported cell preparation method for MS imaging. ToF-SIMS is a highly sensitive, imaging MS technique used to detect and map chemical and molecular information from sample surfaces. ToF-SIMS uses a finely focused (as small as 150 nm), pulsed primary ion beam to desorb secondary ions into a time-of-flight mass spectrometer, creating mass spectral images with excellent spatial resolution and good mass resolution. Each image consists of 256×256 pixels where each pixel is a complete mass spectrum. As with all analytical techniques, sample preparation is critical for reproducible and meaningful IMS results. With any mass spectral technique, to obtain high-quality images of individual cells the cells must be attached to a suitable substrate, free of contamination or interfering components, and accessible to the ionization source. Traditionally for IMS analysis, cells are grown on a conductive substrate and freeze fractured prior to analysis.

Preparation of Single Cells

255

The original freeze–fracture technique was reported in 1957 and has been widely used to prepare cells for membrane analysis by electron microscopy (22, 23). Working from this model, Chandra et al. developed a modified freeze–fracture method to prepare cells for imaging MS analysis (24). Currently, the majority of cellular MS imaging studies utilize some variation of this method. Improvements to Chandra’s technique, most notably preservation of cryogenic temperatures throughout the fracturing and analysis (25) and freeze-etching with vapor matrix deposition (26) have resulted in improvements to cellular imaging ((14) for example). Freeze–fracture methods, however, have several disadvantages: they require cryogenic facilities, generate low yields of suitably fractured cells, and, by design, tend to fracture cells between the leaflets of the membrane bilayer. Because the fracture plane is generally inside the membrane bilayer, the cytoplasm of the cell remains obscured by a layer of phospholipids. To circumvent some of these disadvantages, several groups have reported other cellular preparation approaches for IMS. Nygren et al. (27) freeze-dried cells in ammonium formate and then imprinted on silver foil, thus collecting only the membrane lipids for analysis. Liu et al. (19) fix cells in 70% ethanol, preserving the morphology of the cells but not the chemical contents of the cytoplasm. Parry and Winograd have reported embedding cells in a trehalose and glycerol matrix, followed by freeze-drying (28). Altelaar et al. (29) used sucrose and water washes while Fletcher et al. (15) utilized only a water wash; both groups followed these washes with freeze-drying. This chapter provides an in-depth protocol for our recently reported cellular preparation procedure: a simple “wash and dry” procedure using an iso-osmotic ammonium acetate solution and a gentle argon drying procedure (30). This method sufficiently cleans individual cells for unobscured mass spectral analysis while at the same time preserving the molecular content of the cells (30). We have found that both the wash and dry procedures are critical to the quality of the mass spectral images produced. Components from cell-culture growth medium, most conspicuously salts, seriously interfere with spectral quality making it nearly impossible to collect meaningful molecular information. The simple “wash and dry” cellular preparation technique detailed below not only successfully removes interferences from the growth medium but also permits delicate cells to remain intact until just before the cells are completely dry in order to obtain the maximum molecular information from each cell. Furthermore, this technique is applicable to a wide variety of cell types, generates reproducible results over an extended time frame, alleviates the need for cryogenic facilities, and produces a high yield of cells suitable for mass spectrometric analysis (30). Importantly, the

256

Berman, Fortson, and Kulp

procedure also allows for the imaging and profiling of both membrane and cytosolic molecular contents, as evidenced by excellent localization of both phosphocholine and potassium in the mass spectral images (30). This detailed preparation technique allows routine imaging MS analysis of individual cells, making possible a wide variety of experiments to further scientific understanding of molecular processes within individual cells.

2. Materials 2.1. Silicon Imaging Substrates

1. Standard 6-inch silicon wafers (University Wafer, South Boston, MA, USA). 2. Diamond-tipped scribe. 3. 100% ethanol and low-lint wiper (Kimwipe). 4. Glass or PFA (perfluoroalkoxy) storage and transfer containers (see Note 1).

2.2. Cell Culture

1. MCF7 human breast adenocarcinoma cells, NIH/3T3 mouse fibroblasts, and MDCK canine kidney cells (American Type Culture Collection (ATCC), Manassas, VA, USA). 2. Growth medium for MCF7 Cells: Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 5% fetal bovine serum (FBS), 1% non-essential amino acids, 10 µg/ml insulin, 2 mM L-glutamine, and 1% penicillin/ streptomycin (see Note 2). 3. Growth medium for NIH/3T3 mouse fibroblasts: DMEM supplemented with 4 mM L-glutamine, adjusted to contain 1.5 g/l sodium bicarbonate, 4.5 g/l glucose, and 10% FBS (see Note 2). 4. Growth medium for MDCK canine kidney cells: Eagle minimum essential medium (MEM) with 2 mM L-glutamine, adjusted to contain 1.5 g/l sodium bicarbonate, 0.1 mM non-essential amino acids, 1.0 mM sodium pyruvate, and 10% FBS (see Note 2). 5. Hank’s balanced salts solution (HBSS) (Sigma-Aldrich, St. Louis, MO, USA). 6. 0.25% (w/v) trypsin (Gibco, Invitrogen) diluted in Ca, Mg-free phosphate buffered saline (PBS). 7. Cell-culture flasks, 75 cm3 canted neck (VWR Scientific, West Chester, PA, USA).

Preparation of Single Cells

257

8. 60-mm glass petri dishes, autoclaved before each use (see Note 1). 2.3. Cell Proliferation

1. Standard 96-well plates for cell culture. 2. Aqueous Non-Radioactive Cell Proliferation Assay (Promega, Madison WI, USA): Briefly, this colormetric assay measures the bioreduction of a tetrazolium compound (3(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2(4-sulfophenyl)-2H-tetrazolium, inner salt; MTS) to a formazan product. The conversion of MTS is directly proportional to the number of living cells in culture. 3. A standard multi-well plate reader (we use BIO-RAD Model 680 Microplate Reader) with absorbance measurement capability at 490 nm.

2.4. Washing and Drying

1. Ammonium acetate (Sigma, St. Louis, MO, USA): 150 mM in Millipore Milli-Q water (18.2 M cm) with pH adjusted to 7.5 using a 1:9 solution of phosphoric acid and 1 M R ammonium hydroxide and sterile filtered using a Stericup Vacuum Filter Cup (Millipore Corporation, Billerica, MA, USA). Store at room temperature under sterile conditions (see Note 3). 2. Argon gas with attached needle valve to control argon flow rate and slit nozzle of approximately 1×2 mm (see Fig. 15.1 and Note 4).

Fig. 15.1. Photographic illustration of the experimental setup for the presented “wash and dry” procedure which produces cells suitably cleaned for imaging mass spectrometry analysis. An ammonium acetate washing solution is quickly removed from cells on a sample substrate using a gentle argon flow.

258

Berman, Fortson, and Kulp

3. Methods 3.1. Silicon Imaging Substrates

1. Standard silicon wafers are cut into approximately 1.2 cm square silicon substrates. 2. It is helpful to mark, using a diamond scribe, each substrate with a distinctive serial number or other marking on the back (non-mirror-finish) side to facilitate sample identification. 3. Each substrate is cleaned on both sides with 100% ethanol using a low-lint wiper. 4. Store substrates in glass or PFA containers to avoid contamination with silicones and other plastic components (see Note 1). 5. Just prior to cell-culture use, substrates are sterilized by UV irradiation overnight.

3.2. Cell Culture

1. MCF7 human breast adenocarcinoma cells are grown in the MCF7 DMEM growth medium at 37◦ C with 5% CO2 . 2. NIH/3T3 mouse fibroblasts are grown in theNIH/3T3 DMEM growth medium at 37◦ C with 5% CO2 . 3. MDCK canine kidney cells are grown in the MDCK MEM growth medium at 37◦ C with 5% CO2 . 4. For all cell types, cells are grown in cell-culture flasks and passaged when approaching 75% confluence to create new stock cultures and experimental cultures (see below). 5. Cell passaging consists of removing and discarding cell medium followed by washing with 10 ml Hank’s balanced salt solution. Cells are then treated with 2.0–3.0 ml trypsin solution and incubated until the cell layer is dispersed. The trypsin is removed and discarded and cells are then split 1–4 into 25 ml of fresh growth medium for each cell type. 6. Cell cultures are not permitted to exceed 15 passages to avoid morphological changes in the cells. 7. Experimental cultures are created by plating approximately 15,000 cells in 60-mm glass petri dishes containing four sterile silicon substrates. Cells are allowed to attach overnight on the polished side of the silicon substrates at 37◦ C with 5% CO2 .

3.3. Cell Proliferation

1. The cellular proliferation assay is used to measure the effects of various washing solutions on cell viability (see Note 5). 2. Cells are plated at 2×104 cells/well in a 96-well plate and grown overnight at 37◦ C, 5% CO2 .

Preparation of Single Cells

259

3. In each plate, one row is used per experimental condition, ensuring eight replicates per plate per experiment. Experiments are performed on at least five separate plates to provide sufficient statistical significance for observed differences in cell proliferation. 4. In each plate, at least one row is reserved for the positive control. These cells are treated by removing medium and replacing with fresh medium. This ensures that any loss of cells or viability caused simply by the change of medium is accounted for. 5. In each plate, at least one row is reserved for the negative control. These cells are treated by removing growth medium, adding 100 µl of 70% ethanol, immediately removing the ethanol, and adding fresh cell medium to the wells. 6. Growth medium is removed from cells in the experimental wells and 100 µl of a washing solution is added and immediately removed. Fresh cell medium is then returned to the wells. 7. The plate is then placed back into the incubator for 48 h to allow for cell growth. 8. Cell proliferation is quantified by using the Aqueous NonRadioactive Cell Proliferation Assay according to the manufacturer’s instructions and absorbance of the wells is measured using a standard multi-well plate reader. 9. Statistical significance of differences in cell growth is tested using a one-sample Student’s t-test (each condition compared to control) and a two-tailed t-test assuming unequal variances (different washing solutions compared to each other). 10. Figure 15.2 shows the results of comparing MCF7 proliferation in the suggested ammonium acetate washing solution with water and the 70% ethanol negative control (see Note 6). 3.4. Washing and Drying

1. The setup for the washing and drying procedure is shown in Fig. 15.1. 2. An individual silicon substrate is carefully removed from the petri dish in which the cells have been growing on the silicon substrate with tweezers holding one corner of the substrate. 3. The sample substrate is quickly (=3) ions in peptide spectra. J Am Soc Mass Spectrom, 7, 233–242. 4. Biemann, K. (1990) Sequencing of peptides by tandem mass spectrometry and high-

energy collision-induced dissociation. Methods Enzymol, 193, 455–479. 5. Roth, K. D., Huang, Z. H., Sadagopan, N., Watson, J. T. (1998) Charge derivatization of peptides for analysis by mass spectrometry. Mass Spectrom Rev, 17, 255–274. 6. Keough, T., Youngquist, R. S., Lacey, M. P. (1999) A method for high-sensitivity peptide sequencing using postsource decay matrixassisted laser desorption ionization mass spectrometry. Proc Natl Acad Sci U S A, 96, 7131–7136. 7. Keough, T., Lacey, M. P., Youngquist, R. S. (2000) Derivatization procedures to facilitate de novo sequencing of lysine-terminated tryptic peptides using postsource decay matrix-assisted laser desorption/ionization

338

8.

9.

10.

11.

12.

13.

14.

Franck et al. mass spectrometry. Rapid Commun Mass Spectrom, 14, 2348–2356. Samyn, B., Debyser, G., Sergeant, K., Devreese, B., Van Beeumen, J. (2004) A case study of de novo sequence analysis of Nsulfonated peptides by MALDI TOF/TOF mass spectrometry. J Am Soc Mass Spectrom, 15, 1838–1852. Keough, T., Youngquist, R. S., Lacey, M. P. (2003) Sulfonic acid derivatives for peptide sequencing by MALDI MS. Anal Chem, 75, 156A–165A. Keough, T., Lacey, M. P., Strife, R. J. (2001) Atmospheric pressure matrix-assisted laser desorption/ionization ion trap mass spectrometry of sulfonic acid derivatized tryptic peptides. Rapid Commun Mass Spectrom, 15, 2227–2239. Lee, Y. H., Kim, M. S., Choie, W. S., Min, H. K., Lee, S. W. (2004) Highly informative proteome analysis by combining improved N-terminal sulfonation for de novo peptide sequencing and online capillary reverse-phase liquid chromatography/tandem mass spectrometry. Proteomics, 4, 1684–1694. Lee, Y. H., Han, H., Chang, S. B., Lee, S. W. (2004) Isotope-coded N-terminal sulfonation of peptides allows quantitative proteomic analysis with increased de novo peptide sequencing capability. Rapid Commun Mass Spectrom, 18, 3019–3027. Gevaert, K., Demol, H., Martens, L., Hoorelbeke, B., Puype, M., Goethals, M., Van Damme, J., De Boeck, S., Vandekerckhove, J. (2001) Protein identification based on matrix assisted laser desorption/ionization-post source decaymass spectrometry. Electrophoresis, 22, 1645– 1651. Marekov, L. N., Steinert, P. M. (2003) Charge derivatization by 4-sulfophenyl isothiocyanate enhances peptide sequencing by post-source decay matrix-assisted laser des-

15.

16.

17.

18.

19.

20.

21.

orption/ionization time-of-flight mass spectrometry. J Mass Spectrom, 38, 373–377. Wang, D., Kalb, S. R., Cotter, R. J. (2004) Improved procedures for N-terminal sulfonation of peptides for matrix-assisted laser desorption/ionization post-source decay peptide sequencing. Rapid Commun Mass Spectrom, 18, 96–102. Alley, W. R., Jr., Mechref, Y., Klouckova, I., Novotny, M. V. (2007) Improved collisioninduced dissociation analysis of peptides by matrix-assisted laser desorption/ionization tandem time-of-flight mass spectrometry through 3-sulfobenzoic acid succinimidyl ester labeling. J Proteome Res, 6, 124–132. Huang, Z. H., Wu, J., Roth, K. D., Yang, Y., Gage, D. A., Watson, J. T. (1997) A picomole-scale method for charge derivatization of peptides for sequence analysis by mass spectrometry. Anal Chem, 69, 137–144. Franck, J., El Ayed, M., Wisztorski, M., Salzet, M., Fournier, I. (2009) On tissue N-terminal peptide derivatizations for enhancing proteins identification in bottomup Imaging strategies. Anal Chem, 81, 8305–8317. Lemaire, R., Tabet, J. C., Ducoroy, P., Hendra, J. B., Salzet, M., Fournier, I. (2006) Solid ionic matrixes for direct tissue analysis and MALDI imaging. Anal Chem, 78, 809–819. Lemaire, R., Wisztorski, M., Desmons, A., Tabet, J. C., Day, R., Salzet, M., Fournier, I. (2006) MALDI-MS direct tissue analysis of proteins: improving signal sensitivity using organic treatments. Anal Chem, 78, 7145–7153. Seeley, E. H., Oppenheimer, S. R., Mi, D., Chaurand, P., Caprioli, R. M. (2008) Enhancement of protein sensitivity for MALDI imaging mass spectrometry after chemical treatment of tissue sections. J Am Soc Mass Spectrom, 19, 1069–1077.

Chapter 20 Specific MALDI-MSI: TAG-MASS Jonathan Stauber, Mohamed El Ayed, Maxence Wisztorski, Michel Salzet, and Isabelle Fournier Abstract MALDI imaging as a molecular mass spectrometry imaging technique (MSI) can provide accurate information about molecular composition on a surface. The last decade of MSI development has brought the technology to clinical and biomedical applications as a complementary technique of MRI and other molecular imaging. Then, this IMS technique is used for endogenous and exogenous molecule detection in pharmaceutical and biomedical fields. However, some limitations still exist due to physical and chemical aspects, and sensitivity of certain compounds is very low. Thus, we developed a multiplex technique for fast detection of different compound natures. The multiplex MALDI imaging technique uses a photocleavable group that can be detect easily by MALDI instrument. These techniques of targeted imaging using Tag-Mass molecules allow the multiplex detection of compounds like antibodies or oligonucleotides. Here, we describe how we used this technique to detect huge proteins and mRNA by MALDI imaging in rat brain and in a model for regeneration; the leech. Key words: Matrix-assisted laser desorption/ionization, time-of-fight, mass spectrometry, mass spectrometry imaging, mRNA, antigens, photocleavage.

1. Introduction As an innovative technique, MALDI-IMS is a powerful tool for direct detection and localization of endogenous and exogenous molecules within biological samples (1, 2). The last developments have led to use this technique to obtain the distribution of various compounds such as lipids, drugs, peptides, and proteins within tissue sections (1–4). Non-targeted aspect of MALDI-IMS is one of the big advantages of the technology compared to other imaging techniques as well as strength of MS for structural elucidation. S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_20, © Springer Science+Business Media, LLC 2010

339

340

Stauber et al.

Many successful applications of this technique have been undertaken recently. In particular, MALDI-IMS of lipids, peptides, and proteins for clinical applications by studying pathologies has shown to be a very promising application by providing information on the variations of abundance and localization of markers (3–14). Moreover, biological processes bring into play many different signaling pathways involving various classes of molecules ranging from oligonucleotides, to proteins, peptides, and lipids. In particular, the correlation of mRNA expression with their corresponding protein regulation, or more generally the correlation of transcriptome to proteome, is of special interest for better understanding of biological mechanisms. This is especially an essential aspect when studying pathologies for earlier diagnosis. However, some specific classes of biomolecules such as oligonucleotides or sugars are still non- or hardly accessible to the direct analysis by MALDI, as are also, very hydrophobic proteins, membrane proteins, high mass proteins (>30 kDa), or lower abundance ones. Ideally, oligonucleotides should be directly detected from tissues, although their large size and low abundance in cells added to analytical difficulties in mass spectrometry (salts adducts and gas phase instability) render their analysis difficult (15). In the same time, multiplex techniques are necessary for diagnosis and prognosis. More and more tissue micro arrays (TMAs) are used today to analyze large number of disease tissues and new, fast, and reproducible multiplex techniques are necessary (9). We, thus, have proposed a new concept of possible multiplex and specific detection and tracking of biomolecules with a special focus on mRNA and proteins for transcriptome/proteome correlations. This concept relies on affinity detection by using a specific designed probe, called Tag-Mass which can be detected by mass spectrometry (16). Tag-Mass offers more selectivity to MALDI-MSI for selectively and specifically tracking known markers of physiological stages in cohorts of samples with a high sensitivity (3, 16, 17). 1.1. The Tag-Mass Concept: Selective Multiplexed Imaging of Biomolecules from Tissue Sections

The “Tag-Mass” strategy is an affinity-based strategy where a probe is directed against a specific target, using a probe that can be imaged by MALDI-MSI (3, 16, 17). The Tag-Mass is a modified probe bearing a reporter group where the reporter group is used in MALDI-MSI to indirectly obtain the image of the probe. The reporter is designed to be a molecule of known molecular mass that is easily detectable under MALDI conditions taking care to use a molecule that is not corresponding to an endogenous compound. To image a probe indirectly via the direct image of the reporter, the reporter must be linked to the probe and released in the final step just before or during the MALDI sequence.

Specific MALDI-MSI: TAG-MASS

341

In the Tag-Mass, the release of the reporter group is obtained by photodissociation under the MALDI laser irradiation using a photocleavable moiety that binds the reporter to the probe (Fig. 20.1a). Thus, the reporter is detached from the probe during the MALDI-MSI acquisition. Many different reporters can be used for this purpose, but most of the times peptides were used. The photocleavable linker is chosen to present a specific absorption band in the UV at a wavelength (340 nm) very closed to that of MALDI lasers (i.e., 337–355 nm). Thus, after hybridization of the modified probe to its target, a classical MALDI-MSI sequence is performed. At a specific location of the acquisition, the presence of a probe will be signed by the presence of the reporter released under the MALDI laser irradiation, which traduced by the observation of a peak at the m/z expected for the reporter (Fig. 20.1b). Reconstruction of the reporter molecular image gives, then, the image of the probe, i.e., those of the targeted molecule (3, 16, 17). Such a concept is compatible with all types of probes including mRNA probes, antibody probes, or even lectins or aptamers owing to image with high selectivity, respectively, mRNA, antigens, oligosaccharides (including glycosylated proteins), and drugs. Tag-mass workflow, MALDI-MSI, is combined with hybridization techniques including in situ hybridization (ISH) and immunohistochemistry (IHC) (3, 16, 17). Specific MALDI-MSI or Tag-Mass MSI by using reporter moieties that can be distinguished by their change in m/z is a technique that can be used in multiplex conditions. Theoretically, there are no limits in the multiplexing conditions except the hybridization step itself because of kinetic competition during the affinity reactions or steric obstruction problems. Tag-Mass can also be use for semi-quantification in multiplex conditions by using a reporter presenting the same physico-chemicals properties, i.e., same analytical behavior using, for example, isotopically labeled reporter such as differentially deuterated peptides. This concept can be extended by looking for alternative ways of releasing the reporter moiety, e.g., chemical released or even released by prompt fragmentation pathways (i.e., before the end of the delay time period). The reporter can also be designed to be observable in LDI conditions avoiding, thus, the use of the MALDI matrix (18, 19). Although this latest solution could be less sensitive than using MALDI conditions but would increase spatial resolution of images. Extension to other ion production sources can also be searched. For secondary ion mass spectrometry (20, 21) (SIMS) or laser ablation inductively coupled plasma mass spectrometry (22–24) (LA-ICP), probes bearing directly a monoatomic element easily detectable by these techniques should be used, if the element as a good sensitivity of analysis and is not present naturally in the surface to study. Such techniques induce quite important fragmentation yields and the reporter element

342

Stauber et al.

a

b

TAG-MASS hν Multiplex hybridization

100

Tagged antibody probe Tagged aptamer probe Drug

mRNA

I(%)

Protein

MALDI-MS

0 1641.0

Tagged lectin probe

or

Mass (m/z)

1817.0

MALDI specific Imaging

mRNA oligosaccharides Tagged oligonucleotide probes

Fig. 20.1. (a) Schematic representation of the reporter release under photodissociation by the MALDI laser using a photocleavable-reporter system coupled to the probe. (b) Workflow of multiplex specific MALDI-MSI (Tag-Mass).

will appear as a fragmentation product. For example, gold-labeled secondary antibodies are a good solution for imaging antigens in LA-ICP MS at a spatial resolution below 10 µm. We present, here, the workflows for Tag-Mass of antigens and mRNA using a photocleavable probe bearing a peptide as reporter moiety. For antibodies, preference was given to use indirect IHC with a primary–secondary antibody system. Indeed, indirect IHC is known to present better performances by decreasing steric obstruction problems and increasing detection level, since secondary antibodies will recognize consensus epitope present in the

Specific MALDI-MSI: TAG-MASS

343

primary antibody sequence allowing attachment of several secondary antibodies. Moreover, secondary antibodies are easier to produce since they require much less specificity. For mRNA, modified uracile nucleotides were used. This requires a specific synthesis in order to add the photocleavable group and reporter moiety directly on the nucleotide basis for keeping both 3′ and 5′ termini free. In fact, the modified nucleotide is to be used for the probe amplification. In former experiments, modified primers (by the addition of a photocleavable-reporter system) were used. This approach had revealed several disadvantages including lake of sensitivity (only one reporter per probe), high cost (specific synthesis required for each mRNA to be localized), and impossibility to amplify the probe by in vitro translation (only one terminus of the primer free). Development of modified uracile nucleotides was a great advance in this respect. Modified nucleotides are available for all probes construction, the sensitivity is increased by the incorporation of several reporters in the probe sequence (amplification of the signal) and probes can be obtained by in vitro translation. Only tagged Uracile strategy will be presented here. Specific MALDI-MSI can also be performed in multiplexing conditions. An example of duplex imaging of two antigens (Cystatin B/Cathepsin D) from a FPE tissue section of the leech T. tessulatum are presented here as an example for multiplexing.

2. Materials 2.1. Preparation of Frozen Tissue Sections 2.1.1. Snap-Frozen Tissues

1. Isopentane cooled at –45◦ C with dry ice. Vapors may cause drowsiness and dizziness, so work in a hood.

2.1.2. Tissue Cryosection and Thaw Mounted

1. Optimal cutting temperature polymer, OCT. 2. Indium tin oxide (ITO)-coated glass slides or other holder compatible with mass spectrometry analysis. 3. A cryomicrotome, Leica CM150S (Leica Microsystems, Nanterre, France).

2.1.3. Pre-analysis Treatment: Tissue Fixation

1. Ethanol 75% (–20◦ C): 75 ml of absolute ethanol (≥99.8%) and water (HPLC grade) to 100 ml. Prepare fresh. Store at –20◦ C. 2. Ethanol 95% (−20◦ C): 95 ml of absolute ethanol (≥99.8%) and water (HPLC grade) to 100 ml. Prepare fresh. Store at −20◦ C.

344

Stauber et al.

2.1.4. Pre-analysis Treatment: Lipids Removal

1. Chloroform (−20◦ C): 100 ml of chloroform (≥99.9%). Store at −20◦ C. Chloroform is harmful by inhalation, so work in the hood.

2.2. Preparation of FFPE Tissue Section 2.2.1. FFPE Tissue Section

1. Indium tin oxide (ITO)-coated glass slides or other holder compatible with mass spectrometry. 2. Water: 100 ml of water (HPLC grade). Prepare fresh. 3. A microtome and an hotplate warm at 50◦ C.

2.2.2. FFPE Tissue Dewaxing

1. Xylene: 100 ml of xylene (≥99.9%). Xylene is harmful by inhalation, so work in the hood. 2. Ethanol 100%. Prepare fresh. 3. Ethanol 95%: 95 ml of absolute ethanol (≥99.8%) and water to 100 ml. Prepare fresh. 4. Ethanol 75%: 75 ml of absolute ethanol (≥99.8%) and water to 100 ml. Prepare fresh. 5. Ethanol 30%: 30 ml of absolute ethanol (≥99.8%) and water to 100 ml. Prepare fresh. 6. Water: 100 ml of water (HPLC grade). Prepare fresh.

2.3. Hybridization Buffers for In Situ Hybridization (ISH)

1. Buffer solution glycine 0.1 M/Tris HCl buffer (pH 7.4). 2. RNAse inhibitoring activity solution: Proteinase K (1 µg/µl in 1 M/Tris HCl and 0.5 M EDTA, pH 8). 3. Post-fixation buffer: 4% paraformaldehyde (0.1 M Phosphate/ 5 mM MgCl2 buffer, pH 7.4), 15 min then triethanolamine (0.1 M, pH 8), 10 min. 4. Washing solution: 20× SSC buffer: standard sodium citrate solution : for a 20× SSC solution dissolve 701.28 g NaCl and 352.92 g NaCitrate in a recipe to make 4 l (check to have pH to 7.0) and bring final volume to 4 l , then Autoclave the solution in order to be RNAse free). 5. Dehydration by ethanol (30◦ , 70◦ , 96◦ ). 6. Probes denaturation (100◦ C, 10 min). 7. Hybridization buffer: 0.01 M dextran sulfate, 0.2 M formamide, 20× SSC 20%, 100 × Denhardt’s 10%). 16 h, 55◦ C. 8. Non-hybridized probe degradation buffer: RNase (10 µg/ ml), 37◦ C, 30 min.

Specific MALDI-MSI: TAG-MASS

345

9. Rinsed steps: a. 20 and 10 mM 2-mercaptoethanol solutions, 10 min. b. 0.5× and 0.1× SSC. c. Ultrapure water. 2.4. Hybridization Buffers for Immunocytochemistry (IHC)

1. Incubation buffer: 0.1 M PBS/1% BSA/1% normal goat serum/0.05% Triton X100. 2. Primary antibody incubation, overnight, 4◦ C, on rocking. 3. Washing solution: phosphate-buffered saline (PBS). Prepare 10× stock with 1.37 M NaCl, 27 mM KCl, 100 mM Na2HPO4, 18 mM KH2PO4 (adjust to pH 7.4 with HCl if necessary) and autoclave before storage at room temperature. Prepare working solution by dilution of one part with nine parts water. 4. Secondary antibody solution: antibody diluted in incubation buffer at room temperature, on rocking. Antibody is either anti-rabbit IgG 1/100 developed in goat (Jackson Immunoresearch, Inc., Europe LTD); FITC-conjugated secondary antibody anti-rabbit IgG 1/100 developed in goat (Jackson Immunoresearch, Inc., Europe LTD) or photocleaved tagged antibody 1/100 (Imabiotech, France). 5. Revelation a. Photocleavable tagged antibody, precleavage under UV 5 min Staining substrate for peroxidase antibody in chloronapthol with 0.05% H2 O2 for detection. b. For FITC ICC, slices were prepared using phenylenediamine in glycerol.

2.5. Matrix Deposition for Proteins Analysis 2.5.1. Using a Microspotter

1. SA/ANI solution: 1.5 equivalent of aniline (ANI) were added to a solution containing 40 mg/ ml of sinapinic acid (SA) in acetonitrile/aqueous TFA 0.1% (6:4, v/v). Aniline and TFA are toxic, so work in the hood. 2. Chemical Inkjet Printer CHIP-1000 (Shimadzu Biotech, Kyoto, Japan).

2.5.2. Using an Automatic Sprayer

1. SA/ANI solution: 1.5 equivalent of aniline (ANI) were added to a solution containing 40 mg/ ml of sinapinic acid (SA) in acetonitrile/aqueous TFA 0.1% (6:4, v/v). Aniline and TFA are toxic, so work in the hood. 2. ImagePrep (Bruker Daltonics, Bremen, Germany).

346

Stauber et al.

2.6. Matrix Deposition for Peptides Analysis 2.6.1. Using a Microspotter

1. ANI solution: 1.5 equivalent of aniline (ANI) were added to a solution containing 10 mg/ml of α-cyano-4hydroxycinnamic acid (HCCA) in acetonitrile/aqueous TFA 0.1% (6:4, v/v). Aniline and TFA are toxic, so work in the hood. 2. Chemical Inkjet Printer CHIP-1000 (Shimadzu Biotech, Kyoto, Japan).

2.6.2. Using an Automatic Sprayer

1. ANI solution: 1.5 equivalent of aniline (ANI) were added to a solution containing 10 mg/ml of α-cyano-4hydroxycinnamic acid (HCCA) in acetonitrile/aqueous TFA 0.1% (6:4, v/v). Aniline and TFA are toxic, so work in the hood. 2. ImagePrep (Bruker Daltonics, Bremen, Germany).

2.7. Mass Spectrometry Analysis 2.7.1. MALDI-MSI Experiment

1. Peptide calibration standard II (Bruker Daltonics, Wissenbourg, France): Angiotensin II, angiotensin I, substance P, bombesin, ACTH clip 1–17, ACTH clip 18–39, Somatostatin 28, Bradykinin Fragment 1–7, Renin Substrate Tetradecapeptide porcine. Covered mass range: ∼700–3,200 Da. Store at −20◦ C. 2. Protein Calibration Standard I (Bruker Daltonics, Wissenbourg, France): Insulin, ubiquitin I, cytochrome C, myoglobin. Covered mass range: ∼5,000–17,500 Da. Store at −20◦ C. 3. An Ultraflex II TOF–TOF equipped with a Smartbeam laser and all the Flex software suite (FlexControl, FlexAnalysis, and FlexImaging) (Bruker Daltonics, Bremen, Germany).

2.7.2. MS/MS Analysis

1. An Ultraflex II TOF–TOF equipped with a Smartbeam laser and all the Flex software suite (FlexControl, FlexAnalysis, and FlexImaging) (Bruker Daltonics, Bremen, Germany). 2. Biotools (Bruker Daltonics, Bremen, Germany).

2.8. MALDI Imaging Analysis

1. MALDI matrix used for antibody detection: α-Cyano-4hydroxycinnamic acid (HCCA) and 3-hydroxypicolinic acid (3-HPA) (Sigma-Aldrich).

Specific MALDI-MSI: TAG-MASS

347

2. Calibrant solutions of angiotensin II, Des-Arg-bradykinin, substance P, ACTH 18–39, ACTH 7–38, and bovine insulin (Sigma-Aldrich). 3. Trifluoroacetic acid (TFA) (Applied Biosystems). Acetonitrile and methanol (J.T. Baker). 4. Sprayer of matrix (ImagePrep, Bruker Daltonics).

3. Methods The methods are described according to their sequential used in experiments. One part is dedicated to the protocols or hybridization themselves, i.e., ISH and IHC using the photocleavable modified probe for both cryostat sections (from frozen samples or fixed and frozen samples) and microtome sections (for fixed and paraffin-embedded [FPE] samples). It must be noticed that the protocols are tissue and target dependent. This is usual in hybridization. Thus, proposed protocols are given for rat brain tissue sections and shall be slightly modified and optimized for other applications according to the tissue and the probe (specificity, selectivity). Figures 20.2 and 20.3 give examples of MS spectra and molecular images obtained, respectively, for Proenkephalin mRNA imaging and Carboxypeptidase D (CPD) protein imaging, respectively, using the modified dU for the oligonucléotides probe construction and a rabbit secondary modified antibody for CPD in conjunction with a primary antibody directed against rat CPD and raise in rabbits as antibody. The images are reconstructed based on the reporter signal with a peptide close the to bradykinin sequence. Figure 20.2 gives the molecular distribution of Proenkephalin mRNA as obtained by specific MALDI-MSI compared to the localization obtained by classical ISH. A good correlation is observed in between ISH and MALDI-MSI images. Figure 20.3 compared the specific MALDI-MSI localization of CPD with its distribution obtained for classical colorimetric reaction or fluorescence. Specific MALDI-MSI presents higher sensitivity than classical revelation and is not so far from fluorescence detection. Specific MALDI-MSI enables multiplex detection of different epitopes. An example of duplex experiments for proteins is given Fig. 20.4 where Cystatin B/Cathepsin D system is studied in FPE tissue sections of the leech Theromyzon tessulatum. Experiments were performed using two photocleavable secondary antibodies bearing different reporter peptides represented by different m/z. In such an experiment, one primary antibody is raised in rabbit

348

Stauber et al. 100 80

(a) Classical proenkephalin probe

1080.11

3600.6

% Intensity (a.u.)

60 40 20 0 100 80 60

1037.02

3600.6 1072.83 P,[MH]+ 1163.59 (b) U-tagged proenkephalin probe

40 20 0 983.0

1064.6

1146.2

1227.8

1309.4

1391.0

Mass (m/z)

c-image/map of mouse brain section

d-ISH

e-MALDI image

f-picture of rat brain section

Fig. 20.2. Compared MALDI mass spectra in the linear positive mode recorded on two adjacent rat brain sections in the same region of the brain after ISH of double strand oligonucleotide cDNA probe corresponding to proenkephalin for classical untagged proenkephalin probe (a) and the U-tagged proenkephalin probe (b). (f) Corresponding reconstructed MALDI image on the basis of the tag signal obtained by scanning the tissue section after ISH experiment (7,000 spots separated each of 100 µm) compared (e) to proenkephalin mRNA localization in 8-weeks old male C57BL/6 J mouse brain using digoxigenin ISH technique by the Allen Institute (http://www.brainatlas.org/aba). For this experiment, colorimetric detection of bound probe is generated by the alkaline phosphatase substrates nitroblue tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP) that produce a vivid blue/purple particulate reaction product. Figures (a) presents the map/picture representation of the mouse brain and figure (f) the picture of the rat brain section prior to ISH –MALDI imaging experiment. Reprinted from (14) with permission.

with an anti-rabbit photocleavable secondary antibody, whereas the second one is raised in mouse with an anti-mouse photocleavable secondary antibody. The experiment-exemplified multiplex conditions for proteins, although multiplexing, can also be achieved for mRNA.

Specific MALDI-MSI: TAG-MASS

349

1670.36

100

2357.9 (a) FITC secondary antibody

80 60 1698.34

%Intensity (a.u.)

40

1748.59

20 0 100

P-PC [MO+H]+ 1686.43

80 60

1637.7 P-PC MH+ 1703.23

(b) Tagged secondary antibody 1783.88

40 1752.90

20 0

1783.93

1641.0

1676.2

1711.4 1746.6 Mass (m/z)

(c) MALDI CPD protein image

(e) fluorescence

1781.8

1817.0

(d) Tissue section picture

(e) fluorescence

(f) 4-chloronaphtol

Fig. 20.3. Specific MALDI-MSI of CPD protein from a rat brain tissue section performed using a rabbit anti-rat primary antibody directed against CPD and the modified photocleavable goat anti-rabbit secondary antibody. (a–b) MALDI-MS spectra recorder for serial rat brain tissue sections after the IHC experiments for the tagged secondary antibody (b) compared to a FITC secondary antibody (a) for the same IHC conditions. (c) Molecular MALDI images reconstructed using the signal at m/z 1,686.43 of the reporter moiety (d) tissue section image before IHC experiments. (e and f) Comparison with fluorescence and 4-cholonaphaol detection using, respectively, a FITC or preroxydase tagged secondary antibody. Reprinted from (14) with permission.

3.1. Tissue Treatment 3.1.1. Tissue Snap-Frozen

1. The organ is dissected and rinsed with a saline solution suited for the considered tissue to remove blood and other tissue fragment of the surface. Alternative: prior to sacrifice, the animal can be perfused with the saline solution to remove blood inside the organ.

350

Stauber et al.

1 cm

Cystatin B m/z 1569

Cathepsin D Cystatin B m/z 1569 + Cathepsin D m/z 1309 m/z 1309

Fig. 20.4. Duplex specific MALDI-MSI of Cystatin B/Cathepsin D from a FPE tissue section of the leech T. tessulatum performed using rabbit anti-leech Cathepsin D primary antibody/photocleavable goat anti-rabbit secondary antibody and mouse anti-leech Cystatin B primary antibody/photocleavable goat anti-mouse secondary antibody. Molecular images have been reconstructed on the signals of the two reporter peptides, i.e., m/z 1,309 for Cathepsin D and m/z 15.

2. Morphology of the organ needs to be carefully maintained. Thus, the tissue should not be placed in a tube or wrap in an aluminum foil to avoid deformation of the organ (adaptation to the outlines of the container). 3. Snap-freezing procedure is applied for tissue conservation to maintain tissue morphology and to prevent ice crystals formation and cell explosion. In fact, different rate cooling of parts of the organ or direct dipping of the organ into liquid nitrogen leads to the formation of cracks and fragmentation of the tissue. Therefore the use of isopentane cooled at – 45◦ C with dry ice is recommended. Freezing time is dependent on the size of the organ. It is preferable not to use an embedding media. For very small organs or surgical pieces, cutting without embedding material increases deformations and damages of the tissue sections. In such a case, a solution containing non-polymeric compounds, namely 10% gelatin solutions, helps to obtain high-quality tissue section. Tissue is embedded in 10% gelatin directly after dissection and frozen as previously described. 4. After snap freezing, tissue is removed from isopentane and stored at −80◦ C. We heartily recommend not overpassing a storage period of 6 months. Over 6 months storage, variation in the molecular profiles could be observed if no sample stabilization procedure is performed. Preferentially tissues should be analyzed a few days or weeks after snap freezing.

Specific MALDI-MSI: TAG-MASS

3.1.2. Tissue Cryosection and Thaw Mounted

351

1. The use of cryopreservative solutions containing organic polymers such as optimal cutting temperature (OCT) polymer should be restricted to the attachment of the tissue to the sample holder and not used for complete embedding of the tissue. Moreover, all parts of the cryostat in contact with the tissue must be cleaned to prevent any contamination between two different samples or with the polymer containing solution. In the case of contact between the tissue and cryopreservative solutions containing polymers, MS spectra will be dominated by polymer signals such as PEG signal distribution. 2. Tissue is placed in the cryostat during sufficient time before sectioning for slow warming of the sample to the cryostat temperature. If the tissue is too cold, poor-quality sections are obtained. 3. 10 µm thickness tissue sections are cut using cryomicrotome at −20◦ C. Different tissue types may need other temperature settings. 10 µm thickness is optimal. Smaller sections have not enough molecules for extraction and thicker sections may cause problems of conductivity (due to the insulating nature of tissues) and charge effects by charge accumulation at the sample surface during MALDI analysis. Charge effects will decrease spectral quality in axial TOF configuration instruments resulting in a progressive peak shifting toward the high m/z ratio. 4. Collect the tissue sections onto ITO glass slides pre-cooled at −20◦ C. Transfer is performed by applying the cooled ITO slide onto the section. The cuts are thus stick on the cold slide. Adhesion of the frozen sections to the glass slides is obtained by heating putting fingers under the slide or by placing the slide at room temperature. This transfer procedure, contrarily to classical thaw mounting, prevents formation of ice crystals at the surface of the cryostat microtome cutting plate. 5. Care must be taken of air bubbles formation at the surface of the tissue section that may leads to artifacts during MS analysis. 6. Mounted sections are stored in a sealed container at −80◦ C until their use.

3.1.3. Pre-analysis Treatment: Tissue Fixation

1. A closed container store at −80◦ C is warmed at room temperature in a vacuum desiccator to prevent water condensation at the surface of the frozen slide. 2. After complete drying, the ITO slide is washed. Washing steps are optional and dependent on the molecules to be

352

Stauber et al.

analyzed. Careful washing is crucial for conserving spatial localization of molecules. 3. For analysis of small molecules like lipids or drugs, no washing steps are used. For macromolecules analysis like peptides or proteins washing procedures are generally used. Washing is performed by immersing the glass slide softly in ice-cold 75% ethanol during 30 s. No agitation or shake is needed. This step washes out salts, cells fragments, or residual fluids. 4. Take the slide out and remove the excess of liquid around the section. A stream of nitrogen over the surface could help to remove excess of ethanol. 5. The ITO glass slide is then placed in a vacuum desiccator to completely dry of the tissue. The time of drying is dependent to the size of the section. 6. Optional: a second bath of fresh ice-cold 75% ethanol during 30 s followed by a complete drying under vacuum desiccator can be achieved. 7. After complete drying, the sample is dipped into cold 95% ethanol during 30 s. No agitation or shake is needed. This step prevents degradation of proteome by dehydration and fixation of the tissue. 8. The slide is completely dried like in steps 4 and 5. 3.1.4. Pre-analysis Treatment: Lipids Removal

1. After complete drying, immerse the glass slide softly in icecold chloroform (30 s). No agitation or shake is needed. This step removes lipids (especially phospoholipids) present in high concentration in the tissue (components of cell membranes) and may cause signal suppression in MS spectra. 2. Take the slide out and place it in the vacuum desiccator for complete drying of the tissue.

3.2. Preparation of FFPE Tissue Section 3.2.1. FFPE Tissue Section

1. 10 µm thickness FFPE tissue sections are cut using a microtome at room temperature. Paraffin block can be cooled down −20◦ C prior sectioning to facilitate tissue sectioning. 2. Sections are transferred onto a conductive ITO glass slide on top of a water droplet. 3. Glass slide is warmed up on a hotplate to leave the cuts unfolds. 4. Excess of water is removed and glass slide is stored in an incubator at 30◦ C during 20 min for good adherence.

Specific MALDI-MSI: TAG-MASS

353

Subsequently obtained glass slides with FFPE tissue sections can be stored during over months at room temperature. 3.2.2. FFPE Tissue Dewaxing

1. After complete drying, the glass slide is gently dipped into a bath of xylene during 5 min. This procedure is repeated to fold. No agitation or shake is needed. 2. The slide is then washed in stepwise immersion, 5 min duration each, into 100% ethanol twice, 95% ethanol, 75% ethanol, and 30% ethanol for rehydration of tissue sections. 3. The ITO glass slide is placed in the vacuum desiccator for complete drying of the sections.

3.3. Tissue Preparation for In Situ Hybridization (ISH)

1. 10 µm thickness FFPE tissue sections obtained as previously described are used for ISH. 2. Paraffin is removed by using xylene baths (two times, 15 min), and then tissue is hydrated during 5 min in three steps of different mixed ethanol/water baths (96◦ , 70◦ , 30◦ ). 3. Sections were prepared according to classical ISH protocols. Tissues were incubated in glycine buffer, and then treated for 15 min with proteinase K for protein digestion. 4. After post-fixation with 4% paraformaldehyde for 15 min, then a bath with triethanolamine (0.1 M, pH 8) was carried out for 10 min. 5. Sections were washed with 2× SSC, then ultrapure water for 5 min. Probes were denaturized at 100◦ C for 10 min, and after a 3 step tissue dehydration (30◦ , 70◦ , 96◦ ), hybridization was done for 16 h at 55◦ C dissolving cDNA probes in hybridization buffer (Dextran sulfate 10%, formamide 50%, 20× SSC 20%, 100× Denhardt’s 10%). 6. Tissues were incubated with RNase, then rinsed 10 min with successive SSC solutions and twice 0.5×SSC solutions at 55◦ C for 30 min. After rinsing slices with 0.1×SSC for 5 min at room temperature, one bath of ultrapure water was carried out to remove the excess of polymers. Tissues were kept drying at room temperature before MALDI matrix application.

3.4. Tissue Preparation for Immunocytochemistry (ICC)

1. Frozen sections of rat obtained as previously described are used for ICC. 2. They were incubated at room temperature with 500 µl of incubation buffer for 30 min. The same buffer was used to dilute anti-rat Carboxypeptidase D (CPD) antibody (1:400), and incubation was performed overnight at 4◦ C.

354

Stauber et al.

3. For the leech, Cystatin B and Cathepsin D primary antibodies are used at different concentrations (1/500 Cystatin B and 1/400 Cathepsin D). 4. After washing three times in PBS, sections were incubated with peroxidase conjugated secondary antibody or FITC-conjugated secondary antibody or using photocleavable tagged antibody for 80 min at room temperature. 5. After another three washing steps in PBS buffer, the sections for peroxydase ICC were incubated in chloronapthol with 0.05% H2 O2 for detection. Reaction was stopped with several PBS and ultrapure water baths. For FITC ICC, slices were prepared using phenylenediamine in glycerol. For photocleavable tagged antibody, tissues were rinsed three times for 5 min with ultrapure water to remove salts, and sections were kept drying at room temperature in dark before matrix application. Tissues were then compared using microscopy. 3.5. Peptide Reporter Analysis

3.5.1. Peptide Reporter Analysis Using Dry Droplet

For classical analysis, 1 µl of sample solution and 1 µl of matrix solution (HCCA/ANI) were mixed on the MALDI plate using the dried-droplet technique as a standard control for the different Tag-Mass molecules before imaging.

3.5.2. Peptide Reporter Analysis Using Microspotter

1. An ITO slide after washing step for frozen tissues or digestion for FFPE or frozen tissues is used. 2. On each defined spot, 20 nl of HCCA/ANI solution is applied. 5 droplets of 100 pl are deposited at each spot per cycle, then 40 iterations are necessary to obtain the total volume. For slides after digestion, the matrix is deposited with the same array than the one used for trypsin deposition. In this case matrix is deposited exactly at the same position than the trypsin. 3. Check matrix coverage using an optical microscope. 4. A rapid MS analysis on one spot is recommended to verify that a sufficient amount of matrix is deposited. Increase of iteration number may improve MSI when signal intensity appears to be low.

3.5.3. Peptide Reporter Analysis Using an Automatic Sprayer

1. An ITO slide after washing step for frozen tissue or digestion for FFPE or frozen tissues is used.

Specific MALDI-MSI: TAG-MASS

355

2. A method with different step of spraying, incubation, and drying phase is needed. The ImagePrep method for HCCA/ANI deposition is based on the normal HCCA method included in the ImagePrep. Optimization is required for each type of tissue. Briefly, the spray time is around 2 s (depending the surface of tissue section). An incubation time of 20 s (except for initialization phase: 10 s) allows an effective extraction of proteins. A particular attention is drawn to correctly set the drying time for complete crystallization on the tissue section. If the time is too short, the section will be too wet and a delocalization of molecules will be observed. The minimum drying time is around 120 s. 3. Check matrix coverage using an optical microscope. 4. A rapid MS analysis at one position can be performed to check out that a sufficient amount of matrix has been deposited. If not, some cycles of the last phase of deposition can be done again and may improve MSI when signals intensity is too low. 3.6. Mass Spectrometry Analysis 3.6.1. MALDI-MSI Experiment: In Linear Mode

3.6.2. Mass Spectrometry Analysis for Proteins MSI (For Frozen Tissue Analysis Exclusively)

Acquisition parameters were set to acceleration voltage, 20 kV; first grid voltage, 94%; guide-wire voltage, 0.05%; extraction delay time, 200 ns. Each spectrum was an average of 500 laser shots at 100 Hz. 1. 0.5 µl of protein calibration solution is deposited near to the tissue section and mix with 0.5 µl of HCCA /ANI solution. 2. The mass spectrometer is calibrated with the calibration spot. 3. Using FlexImaging an area of interest is selected on the tissue after definition of the teaching points. 4. The distance between each measurement point is set. Distance between measurement points is, depending on the method, used for matrix deposition: With Chip-1000 deposition, the spots are generally spaced by 250 µm center to center. It is possible to define the same raster than the one defined during matrix deposition. Due to the size of the spot it is possible to accumulate spectra at different position in the same spot. This increase statistics and reduce spot-to-spot variability.

356

Stauber et al.

With ImagePrep deposition, distance between two measurements can be chosen by the user. Generally the resolution is around 100 µm. 5. In FlexControl, the adequate methods for proteins analysis is set in positive linear mode and a total of 500 spectra are acquired at each position at a laser frequency of 100 Hz. 6. The images are saved and reconstructed using FlexImaging 2.1. 3.6.3. Mass Spectrometry Analysis for Peptides MSI

1. 0.5 µl of peptide calibration solution is deposited near to the tissue section and mixed with 0.5 µl of ANI solution. 2. The mass spectrometer is calibrated with the calibration spot. 3. Using FlexImaging an area of interest is selected on the tissue after definition of the teaching points. 4. The distance between each measurement point is set. Distance between measurement points is dependent of the method used for matrix deposition. 4.1. With Chip-1000, deposition spots are generally spaced by 250 µm center to center. It is possible to define the same raster than for matrix deposition. Due to the size of the spots spectra can be accumulated at different positions in the same spot. 4.2 With ImagePrep deposition, distance between two measurements is chosen by users. Generally the resolution is around 100 µm. 5. In FlexControl, the adequate methods for peptides analysis is set in positive reflector mode and a total of 500 spectra are acquired at each position at a laser frequency of 100 Hz. Although negative reflectron mode can also be used for specific class of peptides. 6. The images are saved and reconstructed using FlexImaging 2.1.

4. Notes 4.1. Photocleavable Tagged Oligonucleotide

The peptide is synthesized on Symphony (Protein Technologies Inc.) and purified on a Delta-Pak C18 15 µm 100A column (Waters). The oligonucleotide is synthesized from 3′ to 5′ on Expedite (Applied BioSystems). The amine function with photocleavable linker is added in 5′ before cleavage and deprotection. These steps are performed using a NH4 OH 28% solution during 24 h in the dark. The amino oligonucleotide is then purified

Specific MALDI-MSI: TAG-MASS

357

on a Delta-Pak C18 15 µm 300A column (Waters). The amino function of the oligonucleotide is coupled with a heterobifunctional reagent comprising a maleimide function. The maleimido oligonucleotide is solubilized in water and added to a 1.2 equivalent of peptide in solution. The mixture is stirred for 16 h. The oligo-peptide conjugate is then purified on a DeltaPak C18 15 µm 300A column (Waters) and characterized by MALDI-MS (Voyager STR, Applied BioSystems). 4.2. Photocleavable Tagged Antibody

Peptides were custom made by Eurogentec S.A. using solid phase peptide synthesis (SPPS) on a 0.25 mmol (millimole) scale using Fmoc (9-fluorenylmethyloxycarbonyl amino-terminus protection) standard synthesis protocols (4 equivalents of FmocAA) with double-coupling reactions (twice 40 min) using TBTU/NMM as activator on a Symphony (Rainin Instrument Co, Woburn, MA, USA) synthesizer. The photocleavable linker (4 equivalents) was introduced manually using DIPCDI/DIPEA (2 h) as activator. Purifications were performed by RP-HPLC on a Waters (Milford, MA, USA) Delta-Pak C18 (15 µm−100A−25×100 mm) column using a Waters liquid chromatography system consisting of Model 600 solvent delivery pump, a Rheodine injector, and a automated gradient controller (solvent A: H2 O-0.125% TFA; solvent B: CH3CN-0.1% TFA, gradients: 5–15 to 30−60% B in 20 min). Detection was carried out using Model M2487 variable wavelength UV detector connected to the Waters Millennium software control unit. The Quality Control was performed by analytical RP-HPLC on a Waters Delta-Pak C18 (5 µm−100A−150×3.9 mm) column (solvent A: H2 O-0.125% TFA; solvent B: CH3CN-0.1% TFA, gradient: 100% A−60% B in 20 min) using a Waters Alliance 2690 Separation Module equipped with a Waters 996 Photodiode Array Detector and by MALDI-TOF MS (Voyager STR, Applied BioSystems). The Functionalization with the photolinker derivatized peptide A was done as follow: a solution of 0.5 mg of MBS in 300 µl of DMF is added to a solution of 4 mg of goat anti-rabbit IgG in 2 ml of PBS and mixed for 30 min. The solution is then desalted on a PD 10 column using 50 mM phosphate buffer at pH =6. To this desalted activated IgG, a solution of 1 mg of the photocleavable derivatized peptide in 300 µl of DMF and 1 ml of PBS is added and stirred for 3 h at room temperature. Afterward, the reaction mixture is dialyzed overnight against PBS (membrane cut-off 12–14,000). In order to prepare this triphosphate, a Fmoc-protected CPG resin was required. The succinylate was prepared from GT115A (100 mg) (Scheme 20.1). The sample was relatively pure but contained a small amount (by TLC) of a higher running nontritylated compound (originates from the Sonogashira reaction

358

Stauber et al. O

H N O

O O HN

H OH N O P

H

O

peptide O

H

NO2

N

O Triphos-O

N

O

H

N

O

OH

O O

NHFmoc

N

HN

H

GT115A O DMTO

N O O

OH O HN

H

CMM660A

N

O DMTO

NHFmoc

N

O

O

O O O

O OH

N

HN

H

CMM653B

N

O DMTO

NHFmoc

O

O

O

O NH-CPG500

Scheme 20.1. Synthesis of a dUTP-peptide conjugates with a photocleavable linker (see text for details).

and does not interfere with subsequent reactions and was not visible in the NMR spectra of the sample). Since it was not possible to purify the succinate, the reaction was modified slightly. It is normal to add two equivalents of succinic anhydride to the reaction to get quantitative yield but if this is not removed completely, the amino residues of the cpg resin can become blocked during functionalization. Therefore, 1.5 equivalents were used since the exact purity of the product is undetermined. The reaction did not go to completion (from TLC this was more than 50%) by comparing the intensity of the components on the TLC by UV (254 nm) and the intensity of the DMT cation on treatment with HCl fumes. Since the non-succinylated product will not react, the resin was functionalized using this mixture. The resin was prepared but the loading is very low, 5.4 µ mol g−1 (180 mg).

Specific MALDI-MSI: TAG-MASS

359

The resin was detritylated using 2% TCA/DCM washed with DCM and the process repeated until no orange color due to the DMT cation was observed. This was then dried (suction under argon) and the resin soaked in pyr/DMF 1:3 (0.4 ml) for 5 min before a solution of 0.1 M Eckstein’s reagent in dioxane was added (0.1 ml). The reaction was allowed to stand for 15 min after which time the resin was washed (dioxane, MeCN) and dried (suction under argon). The resin was then soaked in a solution on 0.5 M bis-(tributylammonium) pyrophosphate in anhydrous DMF and tri-n-butylamine for 20 min and the resin washed (DMF, MeCN) and dried (suction under argon). The product was oxidized (iodine/water/pyridine/THF for 30 min), washed (MeCN), and dried (suction under argon). The Fmoc protecting group was removed (20% piperidine/DMF, 0.5 ml, 20 min) and the resin washed thoroughly (DMF, MeCN) and dried (suction under argon). This was then washed with DCI and a solution of DCI/photolabile amino linker CEP (1:1, 0.5 ml) was added and the reaction was allowed to stand for 20 min. The solution was removed and the resin washed (MeCN) and dried (suction under argon). A mixture of cap A/cap B (1:1, 0.5 ml) was added and the resin soaked for 5 min before removing the capping reagents and washing and drying the resin as before. The product was oxidized (I2 /THF/pyr/H2 O, 5 min) and the resin washed and dried as before. This was cleaved from the resin with cNH4 OH at room temperature for 30 min, then purified by anion exchange HPLC on a Dionex NucleoPac100 HPLC column using the following solvent system Buffer A:0.1 M NH4 Cl with 10% acetonitrile; Buffer B: 1 M NH4 Cl with 10% acetonitrile; flow rate 2.5 ml /min. using 6Triphos.mth. This gave three fractions (A:–7 min, B:–7.9 min, and C:–10.3 min). All three fractions were lyophilized overnight before being desalted by reverse phase HPLC Buffer A: Water; Buffer B: acetonitrile; flow rate 4 ml/min. The three fractions were again lyophilized overnight before being suspended in 200 µl of water. MS showed that CMM661A pk 1 was definitely not the triphosphate but it could be either CMM661pk 2 or 3 (very similar MS profiles). (CMM662A was formed from CMM661A pk 2 and CMM663A was formed from CMM661A pk 3). Both samples were then used in the subsequent reaction. Bicarbonate buffer (10 µl) and the maleimide NHS ester (50 µl) were added to each sample and the reactions agitated overnight. The samples were diluted with milliQ water (500 µl) and filtered. The samples were purified by RP-HPLC, buffer A: 0.1 M TEAA, buffer B: MeCN, flow rate 4 ml /min. using MeCN50.mth and the coupling of the peptide was carried out on these fractions. The use of a cryopreservative solution containing polymer compounds such as a solution with an optimal cutting temperature (OCT) polymer should be restricted to the attachment of

360

Stauber et al.

the tissue to the sample holder and not for wholly embedded the tissue. Moreover, all parts of the cryostat in contact with the tissue need to be cleaned to prevent any contamination between two different samples or with a polymer contain solution. In the case of contact between the tissue and a polymer containing cryopreservative solution, MS spectra will be dominated by polymer signals.

Acknowledgments This study was supported by grants from the Centre National de la Recherche Scientifique (DPI), Ministère de L’Enseignement Supérieur et de la Recherche, the Agence Nationale de la Recherche PCV (To IF) References 1. Caprioli, R. M. (2008) Perspectives on imaging mass spectrometry in biology and medicine. Proteomics, 8, 3679–3680. 2. Cornett, D. S., Reyzer, M. L., Chaurand, P., Caprioli, R. M. (2007) MALDI imaging mass spectrometry: molecular snapshots of biochemical systems. Nat Methods 4, 828–833. 3. Franck, J., Arafah, K., Elayed, M., Bonnel, D., Vergara, D., Jacquet, A., Vinatier, D., Wisztorski, M., Day, R., Fournier, I., Salzet, M. (2009) MALDI imaging: state of the art technology in clinical proteomics. Mol Cell Proteomics, 8, 2023–2033. 4. Murphy, R. C., Hankin, J. A., Barkley, R. M. (2009) Imaging of lipid species by MALDI mass spectrometry. J Lipid Res, 50 Suppl, S317–S322. 5. Grey, A. C., Chaurand, P., Caprioli, R. M., Schey, K. L. (2009) MALDI imaging mass spectrometry of integral membrane proteins from ocular lens and retinal tissue (dagger). J Proteome Res, 8, 3278–3283. 6. Dekker, L. J., van Kampen, J. J., Reedijk, M. L., Burgers, P. C., Gruters, R. A., Osterhaus, A. D., Luider, T. M. (2009) A mass spectrometry based imaging method developed for the intracellular detection of HIV protease inhibitors. Rapid Commun Mass Spectrom, 23, 1183–1188. 7. Wolthuis, R., Travo, A., Nicolet, C., Neuville, A., Gaub, M. P., Guenot, D., Ly, E., Manfait, M., Jeannesson, P., Piot, O. (2008) IR spectral imaging for histopatho-

8.

9.

10.

11.

12.

13.

logical characterization of xenografted human colon carcinomas. Anal Chem, 80, 8461–8469. Wisztorski, M., Croix, D., Macagno, E., Fournier, I., Salzet, M. (2008) Molecular MALDI imaging: an emerging technology for neuroscience studies. Dev Neurobiol, 68, 845–858. Walch, A., Rauser, S., Deininger, S. O., Hofler, H. (2008) MALDI imaging mass spectrometry for direct tissue analysis: a new frontier for molecular histology. Histochem Cell Biol, 130, 421–434. Stauber, J., Lemaire, R., Franck, J., Bonnel, D., Croix, D., Day, R., Wisztorski, M., Fournier, I., Salzet, M. (2008) MALDI imaging of formalin-fixed paraffin-embedded tissues: application to model animals of Parkinson disease for biomarker hunting. J Proteome Res, 7, 969–978. Fournier, I., Wisztorski, M., Salzet, M. (2008) Tissue imaging using MALDI-MS: a new frontier of histopathology proteomics. Expert Rev Proteomics, 5, 413–424. Chaurand, P., Rahman, M. A., Hunt, T., Mobley, J. A., Gu, G., Latham, J. C., Caprioli, R. M., Kasper, S. (2008) Monitoring mouse prostate development by profiling and imaging mass spectrometry. Mol Cell Proteomics 7, 411–423. Burnum, K. E., Tranguch, S., Mi, D., Daikoku, T., Dey, S. K., Caprioli, R. M. (2008) Imaging mass spectrometry reveals unique protein profiles during

Specific MALDI-MSI: TAG-MASS

14.

15.

16.

17.

18.

embryo implantation. Endocrinology, 149, 3274–3278. Lemaire, R., Menguellet, S. A., Stauber, J., Marchaudon, V., Lucot, J. P., Collinet, P., Farine, M. O., Vinatier, D., Day, R., Ducoroy, P., Salzet, M., Fournier, I. (2007) Specific MALDI imaging and profiling for biomarker hunting and validation: fragment of the 11S proteasome activator complex, Reg alpha fragment, is a new potential ovary cancer biomarker. J Proteome Res, 6, 4127–4134. Nordhoff, E., Kirpekar, F., Karas, M., Cramer, R., Hahner, S., Hillenkamp, F., Kristiansen, K., Roepstroff, P., Lezius, A. (1994) Comparison of IR- and UV-matrix-assisted laser desorption/ionization mass spectrometry of oligodeoxynucleotides. Nucleic Acids Res, 22, 2460–2465. Lemaire, R., Stauber, J., Wisztorski, M., Van Camp, C., Desmons, A., Deschamps, M., Proess, G., Rudlof, I., Woods, A. S., Day, R., Salzet, M., Fournier, I. (2007) Tag-mass: specific molecular imaging of transcriptome and proteome by mass spectrometry based on photocleavable tag. J Proteome Res, 6, 2057–2067. Stauber, J., Lemaire, R., Wisztorski, M., Ait-Menguellet, S., Lucot, J.P., Vinatier, D., Desmons, A., Deschamps, M., Proess, G., Rudolf, I., Salzet, M., Fournier, I. (2006) New developments in MALDI imaging mass spectrometry for pathological proteomic studies; introduction to a novel concept, the specific MALDI imaging. Mol Cell Proteomics, 5, S247–S49. Thiery, G., Anselmi, E., Audebourg, A., Darii, E., Abarbri, M., Terris, B., Tabet, J.

19.

20.

21.

22.

23.

24.

361

C., Gut, I. G. (2008) Improvements of TArgeted multiplex mass spectrometry imaging. Proteomics, 8, 3725–3734. Thiery, G., Shchepinov, M. S., Southern, E. M., Audebourg, A., Audard, V., Terris, B., Gut, I. G. (2007) Multiplex target protein imaging in tissue sections by mass spectrometry–TAMSIM. Rapid Commun Mass Spectrom, 21, 823–829. Brunelle, A., Laprevote, O. (2009) Lipid imaging with cluster time-of-flight secondary ion mass spectrometry. Anal Bioanal Chem, 393, 31–35. Touboul, D., Halgand, F., Brunelle, A., Kersting, R., Tallarek, E., Hagenhoff, B., Laprevote, O. (2004) Tissue molecular ion imaging by gold cluster ion bombardment. Anal Chem, 76, 1550–1559. Becker, J. S., Zoriy, M., Matusch, A., Wu, B., Salber, D., Palm, C. (2009) Bioimaging of metals by laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS). Mass Spectrom Rev, 29, 156–175 Becker, J. S., Dobrowolska, J., Zoriy, M., Matusch, A. (2008) Imaging of uranium on rat brain sections using laser ablation inductively coupled plasma mass spectrometry: a new tool for the study of critical substructures affined to heavy metals in tissues. Rapid Commun Mass Spectrom, 22, 2768–2772. Becker, J. S., Zoriy, M. V., Dobrowolska, J., Matucsh, A. (2007) Imaging mass spectrometry in biological tissues by laser ablation inductively coupled plasma mass spectrometry. Eur J Mass Spectrom (Chichester, Eng), 13, 1–6.

Chapter 21 Structurally Selective Imaging Mass Spectrometry by Imaging Ion Mobility-Mass Spectrometry John A. McLean, Larissa S. Fenn, and Jeffrey R. Enders Abstract This chapter describes the utility of structurally based separations combined with imaging mass spectrometry (MS) by ion mobility-MS (IM-MS) approaches. The unique capabilities of combining rapid (µs-ms) IM separations with imaging MS are detailed for an audience ranging from new to potential practitioners in IM-MS technology. Importantly, imaging IM-MS provides the ability to rapidly separate and elucidate various types of endogenous and exogenous biomolecules (e.g., nucleotides, carbohydrates, peptides, and lipids), including isobaric species. Drift tube and traveling wave IM-MS instrumentation are described and specific protocols are presented for calculating ion–neutral collision cross sections (i.e., apparent ion surface area or structure) from experimentally obtained IM-MS data. Special emphasis is placed on the use of imaging IM-MS for the analysis of samples in life sciences research (e.g., thin tissue sections), including selective imaging for peptide/protein and lipid distributions. Future directions for rapid and multiplexed imaging IM-MS/MS are detailed. Key words: Ion mobility, ion mobility-mass spectrometry, IM-MS, imaging mass spectrometry, IMS, MSI, imaging ion mobility-mass spectrometry, structural separations, MALDI, IM-MS/MS.

1. Introduction One of the recent advances in mass spectrometry (MS) instrumentation is the incorporation of post-ionization separations on the basis of ion mobility (IM) combined with subsequent MS analysis (IM-MS). Importantly, IM-MS adds an additional dimension of separations on the basis of analyte structure to facilitate interpretation of MS spectra directly from complex biological samples. Typically separations in the IM dimension are completed in 100 s of microseconds to milliseconds, thus imaging S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_21, © Springer Science+Business Media, LLC 2010

363

364

McLean, Fenn, and Enders

matrix-assisted laser desorption/ionization (MALDI)-IM-MS is performed on the same timescale as contemporary imaging MALDI-MS experiments. In practice, the combination of imaging IM-MS can be thought of as rapid gas-phase electrophoresis at each spatial dimension yielding a 5D data set (i.e., mobility, m/z, relative abundance, and x, y spatial coordinates). This information can be supplemented by performing fragmentation studies at each spatial position in imaging IM-MS/MS experiments (Section 1.4.1). Furthermore, IM gas-phase separations yield direct structural information (ion–neutral collision cross sections, or apparent ion surface area (Å2 )) for analytes that can be interpreted by complementary molecular simulations using either ab initio and/or molecular dynamic techniques. The main focus of this chapter is to illustrate the advantages of merging IM-MS with imaging MALDI-MS, which is aimed primarily at new or potential users of imaging MALDI-IM-MS. The ability to separate molecules based on collision cross section allows for the simplification and validation of complex spectra, such as those commonly encountered in biological imaging MS. The recent commercial availability of IM-MS instruments should ultimately result in this technology becoming a staple in many structural MS and biological laboratories. In order to avoid repetition with other works in this edition, this chapter focuses mainly on the theory of IM-MS separations and the benefits of using this technique for biological imaging experiments. The selected examples described and illustrated herein are intended to underscore the advantages and limitations of imaging MALDI-IM-MS rather than being a comprehensive review of IM-MS research. The reader is directed to several recent reviews for a more detailed description of IM-MS (1–5). 1.1. Ion Mobility Applications to the Life Sciences

Although gas-phase IM separations have existed for well over a century (6) and coupling IM–MS has existed since the early 1960s (7, 8), the utility of IM-MS for biomolecular separations was not fully realized until combined with soft ionization techniques, such as electrospray ionization (ESI) and MALDI (9, 10). The first applications of IM-MS to determine peptide and protein structures were performed in the late 1990s (11–13). Subsequent to these pioneering studies, research over the past decade has extended IM-MS techniques to the study of complex biological samples, such as whole cell lysates (14), plasma (15–18), homogenized tissue (14, 19, 20), non-covalent complexes (21–23), or directly from thin tissue sections (24, 25). However, until very recently, IM-MS was essentially available in only a limited number of laboratories where custom instruments were constructed. The recent introduction of commercially available IM-MS instrumentation, in several forms, has further fueled the integration of IM-MS techniques into life sciences research programs. The following sections describe the theory of

Imaging Ion Mobility-Mass Spectrometry

365

IM separations (Section 1.2), an overview of IM-MS instrumentation (Section 1.3), and data interpretation in IM-MS conformation space (Section 1.4). Materials, methods, and protocols for performing imaging IM-MS of complex biological samples are then detailed (Sections 2 and 3). 1.2. Overview and Theory of Ion Mobility Separations

Ion mobility-mass spectrometers are composed of an ion source, a mobility separation cell, a mass analyzer, and a detector as depicted in Fig. 21.1a. There are many variations to the general

Fig. 21.1. (a) A block diagram of the primary components of biological IM-MS instrumentation. (b) A conceptual depiction of an IM drift cell. A stack of ring electrodes are connected via resistors in series to form a voltage divider, which is typically designed to generate a relatively uniform electrostatic field along the axis of ion propagation. Ions of larger apparent surface area experience more collisions with the neutral drift gas and therefore elute slower than ions of smaller apparent surface area. (c) A hypothetical IM separation for peptide ions exhibiting two distinct structural sub-populations corresponding to globular (left) and to helical (right) conformations. The arrival time distribution data (top axis), or what is measured, can be transformed to a collision cross-section profile (bottom axis) via equation [4] and described in Section 3.1. Adapted with kind permission from Springer Science+Business Media (1) Fig. 1.

366

McLean, Fenn, and Enders

design such as different ion sources (i.e. ESI, MALDI) and types of ion mobility separation cells used (i.e., whether the ions are dispersed in space or time). For imaging IM-MS applications typically time-of-flight (TOF) mass analyzers are used for timescale considerations as described below (see Section 1.3). This chapter focuses on temporal ion dispersion through the use of drift tube or traveling wave ion mobility (DTIM and TWIM, respectively). In contrast with high-energy ion–neutral gas-phase collisions used in collision-induced dissociation (CID), both DTIM and TWIM separations utilize low-energy gas-phase collisions to separate ions on the basis of predominantly molecular surface areas. Briefly, ions are injected into a drift tube filled with a neutral drift gas and migrate under the influence of a weak electrostatic field gradient (Fig. 21.1b). Larger ions have a lower mobility than smaller ions which result in longer drift times versus shorter drift times, respectively. This field is electrostatic for drift tube and electrodynamic for traveling wave separations, respectively. The migration of these ions is impeded by collisions with the neutral drift gas to a degree that is proportional to apparent surface area or collision cross section. Although the experimental parameter obtained from IM separations is the ion arrival time distribution (tatd ) or the time between ion injection and ion detection, it can be converted to collision cross section or apparent surface area as illustrated in Fig. 21.1c. The following description details how this conversion is performed based on the kinetic theory of gases for drift tube separations. For a derivation of ion–neutral collision crosssection theory, the reader is directed to several excellent texts and reviews (26–28). Procedures to estimate collision cross section using traveling wave IM are described elsewhere (29, 30). This section summarizes several of the key equations and practical considerations for determining ion–neutral collision cross sections in uniform electrostatic field IM instrumentation. Further details for experimental implementation will follow in the methods section. 1.2.1. Transforming Drift Time to Collision Cross Section

The separation of ions in a weak electrostatic field (E) is measured as the ion drift velocity (vd ) and is related by the proportionality constant, K, which is the mobility of the ion in a particular neutral gas: vd = KE

[1]

The drift cell is of a fixed length (L), and the velocity of the ion packet is determined by measuring the drift time (td ) of the packet across the drift cell. In evaluating K, the drift velocity of the ion packet depends not only on the electrostatic field strength but also on the pressure (p, torr) of the neutral drift gas and the temperature (T, kelvin) of separation. Therefore, it is conventional practice to report K as the standard or reduced mobility

Imaging Ion Mobility-Mass Spectrometry

367

(K0 ), which normalizes the results to standard temperature and pressure conditions (i.e., 0◦ C and 760 torr): K0 = K

p 273 760 T

[2]

For applications where IM is used to obtain structural information about the ion, such as those in structural proteomics and biophysics, the IM separations are performed using weak electrostatic fields (ca. 20–30 V cm−1 torr−1 ). Provided the field strength is sufficiently weak, or under so-called low-field conditions, a closed equation for the ion–neutral collision cross section can be expressed from the kinetic theory of gases (see Note 1). When the IM separations are performed in low-field conditions, i.e., constant K, the mobility is related to the collision cross section of the ion–neutral pair:   ze 1 1 1/2 760 T 1 1 (18π)1/2 + K0 = 16 mn p 273 N0  (kB T )1/2 mi

[3]

where these parameters include the charge of the ion (ze), the number density of the drift gas at STP (N0 , 2.69×1019 cm−3 ), the reduced mass of the ion–neutral collision pair (ion and neutral masses of mi and mn , respectively), Boltzmann’s constant (kB ), and the ion–neutral collision cross section (). Inspection of equation [3] shows that the mobility of an ion is inversely related to its collision cross section or apparent surface area. Substituting for K0 in equation [3] and rearranging to solve for the collision cross section yields:   (18π)1/2 ze 1 1 1/2 td E 760 T 1 = + 16 mn L p 273 N0 (kB T )1/2 mi

[4]

which is the typical functional form of the equation used to solve for collision cross sections from IM data (see Section 3). Conceptually, the ion–neutral collision cross section can be thought of as the radius of the orientationally averaged projection of the ion in combination with the drift gas, i.e., =π(ri +rHe )2 , as depicted in Fig. 21.2 (27). In both equations [3] and [4], the collisions of ions with neutrals are considered to be a completely elastic process. Thus, the collision cross section obtained is termed the hard-sphere collision cross section. When compared to molecular simulations, these collision cross section measurements can provide detailed structural information about the analyte (31–34). 1.3. Instrumentation Overview

The two time-dispersive methods of IM separation are DTIM and TWIM. Drift tube IM facilitates absolute collision cross section

368

McLean, Fenn, and Enders

Fig. 21.2. Visual representation of the ion–neutral collision cross section measured using DTIM strategies. The radii of the ion (ri ) and the drift gas helium atom (rHe ) can be used to approximate the collision cross section () using the simplified equation  = π (ri +rHe )2 (27).

calculations (35–38). This data can then be compared to molecular simulation results to interpret analyte structural and conformational details. Traveling wave IM utilizes electrodynamic fields, which only provides estimated collision cross sections when measurements are compared to internal standards with previously measured DTIM absolute collision cross sections (29, 30). This is because gas-phase theory is insufficiently developed for the fundamental physical processes in TWIM separations, although recent efforts in this regard have been reported (39). Nevertheless, both DTIM and TWIM instrumentations are increasingly used for imaging IM-MS applications. 1.3.1. Drift Tube Ion Mobility

The first ion mobility instruments were developed on a drift tube design (40). As described in Section 1.2, DTIM-MS instruments are conceptually analogous to typical imaging MALDI-MS instruments with the exception of inserting an IM drift cell between the MALDI source and mass analyzer. Specifically for imaging applications it is important to consider the timescales of separation, which are largely affected by drift cell pressure and length. The fundamental limit to throughput in imaging applications is dependent on the slowest component, which in this case is either the duration of IM separation or the pulse-to-pulse repetition rate of the MALDI laser. Although atmospheric pressure drift cells and reduced pressure (1–10 torr) drift cells have been constructed for IM-MS (41) the latter are much better suited for imaging applications. For example, atmospheric pressure drift cells typically separate ions on a timescale of 10 s of ms (42), whereas reduced pressure drift cells can provide separations in 17,500). As described in Section 1.4.1, CID can be performed in the regions before and after the TWIM drift cell (see Note 2). Generally resolution in the TWIM is carbohydrates>peptides>lipids (1)). Although more pronounced for an m/z range over 2,000, at lower ranges as that depicted in Fig. 21.4b separations are still feasible but the correlations begin to overlap. Nevertheless, imaging IM-MS allows for obtaining more informative images on the basis of structure and m/z whereby isobaric chemical noise is selectively rejected on the basis of structure. The first examples in the literature of combining imaging IMMS was performed using DTIM and demonstrated two important advantages over imaging MS strategies, namely (i) the ability to separate isobaric species on the basis of structure and m/z and (ii) enhanced signal-to-noise ratios by the separation of chemical noise (24, 25). The former is demonstrated through the selective differentiation of nominally isobaric peptide and lipid species as illustrated in Fig. 21.5a. In proof-of-concept experiments, both the lipid and the peptide were spotted onto a thin tissue section of mouse liver (12 µm) using a reagent spotter in either a “\” direction for the peptide or a “/” direction for the lipid (Fig. 21.5a, top left). In the 2D IM-MS plot (Fig. 21.5a, top right), the signal for the protonated forms of the lipid, phosphotidylcholine 34:2, and the peptide, RPPGFSP, overlaps in the m/z spectrum, but is baseline resolved in the IM arrival time distribution with the peptide and lipid centered at 449 and 504 µs, respectively. In Fig. 21.5a bottom, the × represents what would be obtained using conventional imaging MS in the absence of IM, which is a convolution of both the peptide and lipid signals. The right two images are for the same 1 Da mass range (759– 760 m/z), but selectively for structures corresponding to putative peptides “\” and lipids “/,” respectively (25). Thus, structurally selective imaging on the basis of molecular class results in more

◮ Fig. 21.5. (a) Imaging DTIM-MS of a nominally isobaric peptide (RPPGFSP) and lipid (PC 34:2) deposited onto a mouse liver thin tissue section (12 µm) in the pattern of an “×”. The “\” line is RPPGFSP, while the “/” line is a phosphotidylcholine extract, respectively (top left). An optical image of the patterned matrix/analyte spots deposited on the tissue section (top right). A zoomed view in the region of PC 34:2 and RPPGFSP for a representative 2D IM-MS conformation space plot of a mixture of the two analytes. IM-MS signal intensity is indicated by false coloring, where purple and yellow corresponds to the least and most intense signals, respectively. (bottom left) An extracted ion intensity map over the mass range of 759–760 Da representing what would be obtained using conventional imaging MALDI-MS. (bottom

Imaging Ion Mobility-Mass Spectrometry

373

Fig. 21.5. (continued) middle and right) Extracted ion intensity maps for imaging DTIM-MS of the peptide and lipid over the mass range of 759–760 Da and DTIM drift times of 447–451 µs and 502–506 µs, respectively. (b, left) An integrated mass spectrum of cerebrosides directly from rat brain tissue. (b, right) Imaging DTIM-MS of the sodium-coordinated cerebroside 24:0 OH (m/z = 850.7). (b, far right) An optical image of an adjacent rat brain section. Histological abbreviations are Cx – cortex; fmi – forceps minor of the corpus callosum; CPu – caudate putamen (striatum); Acb – nucleus accumbens; ac – anterior commissure; and lo – lateral olfactory tract. Figures (a) and (b) are adapted from (25) and (24), respectively, with permission. Copyright© 2007 Wiley-Liss, Inc.

374

McLean, Fenn, and Enders

accurate images in contrast with conventional imaging MS alone. In Fig. 21.5b, imaging IM-MS is demonstrated for a coronal rat brain section (16 µm) where the image to the right corresponds structurally to lipids and specifically to the sodium-coordinated cerebroside 24:0 OH (m/z = 850.7 Da). This results in enhanced signal to noise for species of interest through the separation of chemical noise and contaminants with IM (24). More recently, imaging IM-MS using a TWIM drift cell has been demonstrated (Fig. 21.6). Although the IM resolution is more limited using TWIM, it is sufficient for the separation of lipids from peptides as illustrated for a coronal thin tissue rat brain

Fig. 21.6. (i) Imaging TWIM-MS data of a rat brain thin tissue section illustrating selective imaging of peptides and lipids on the basis of structure. (ii) Imaging TWIM-MS data obtained in the analysis of a small drug molecule, common name vinblastine, in thin tissue kidney sections of vinblastine-dosed rats. (a) An optical image of the kidney from the whole body section dosed at 6 mg kg−1 IV vinblastine before matrix application. (b) The same tissue section as shown in (a) but imaged by TWIM-MS showing the distribution of vinblastine within the kidney, with the highest intensity (white) showing a broken ring of intensity between the cortex and the medulla. (c) Optical image of the kidney within the whole body section dosed with 3 H vinblastine. (d) Whole body autoradiography of the tissue section shown in (c) indicated is the broken ring of slightly higher intensity (white) between the cortex and the medulla. Reproduced with permission from (52). Copyright 2008 American Chemical Society.

Imaging Ion Mobility-Mass Spectrometry

375

section (Fig. 21.6i, see Note 4). The utility of imaging TWIMMS for mapping the distribution of a drug, vinblastine, in a kidney from a whole body section is shown in Fig. 21.6ii, b. The accuracy of the obtained image is increased through the addition of TWIM due to the removal of isobaric interferences common in highly complex biological samples, such as tissue (52). Since commercial TWIM instrumentation equipped for imaging applications was released in 2008, the present number of reports is rather limited but expected to increase substantially in the near future. 1.4.1. Imaging IM-MS/MS Measurements

When performing imaging experiments, additional confirmation of an unidentified peak is often required. A common practice for increasing confidence in peak assignments is to image in a selected reaction monitoring mode for a fragmentation channel characteristic of the analyte of interest (e.g., using CID (53)). Like traditional MS/MS, the coupled arrangement of IM and MS yields the ability to obtain structure information in the MS dimension by performing IM-MS/MS. In IM-MS/MS mode, MS1 can be accomplished in two ways: (i) time dispersion in the drift tube can perform parent ion selection (2) or (ii) a mass analyzer can be used (49). By placing an ion activation region between the drift tube and mass analyzer ions may be selected for fragmentation according to drift time. Performing the ion selection in this manner provides a multiplex advantage in that all fragment ions will possess the same drift time as the parent ion. This is significant in imaging applications because the sample is limited to the spatial coordinates of a particular pixel. In imaging IM-MS/MS experiments, multiple parent ions can be fragmented whereby fragment ions are correlated to their respective parent ions by drift time. A demonstration of the potential utility of IM-MS/MS is illustrated in Fig. 21.7 for a carbohydrate, lacto-N-fucopentaose 1 (LNFP1). In-source decay fragmentation for this carbohydrate is illustrated in Fig. 21.7a where fragment ions occur at different times in the IM separation. Correlated fragmentation spectra can also be obtained as illustrated in Fig. 21.7b by using both in-source fragmentation and post-IM CID. The CID fragmentation results in ions correlated to the drift time of the parent. Importantly, this results in redundancy of the fragment ions that are observed for higher confidence that particular fragment ions arise from the parent ion of interest. For example, the integrated mass spectra for in-source and CID fragmentation (i.e., pre- and post-IM separation) are illustrated in Fig. 21.7c, d, respectively. When applied to multiple ions, this operation allows for multiple reaction monitoring (MRM) in a single scan. This application is a highly promising yet virtually untapped resource for biomolecular imaging MS, where limited sample exists at each pixel location.

376

McLean, Fenn, and Enders

Fig. 21.7. ESI-TWIM-MS/MS of the carbohydrate, lacto-N-fucopentaose 1 (LNFP1), illustrating two modes of IM-MS/MS. (a) In-source decay fragmentation of LNFP1 followed by TWIM analysis of the fragment ions. (b) In-source decay prior to TWIM separation and collision-induced dissociation following TWIM separation for LNFP1. The latter results in fragment ions to be observed at the same drift time as the parent leading to the possibility for simultaneous CID for various ions at the same time. In both CID and ISD, cross-ring cleavages were seen, but glycosidic bond cleavages were the most abundant type of fragmentation. (c) In-source decay fragmentation spectrum that was extracted from the IM-MS plot above. Along with a zoom in view of the region from ∼500 to 850 Da. Nomenclature for the fragmentation pattern of carbohydrates was first used by Domon and Costello (54). All in-source fragmentation and collision-induced dissociation peaks are labeled utilizing this nomenclature. (d) CID spectrum of the top dotted line for LNFP1 extracted from (b). (c) and (d) can be compared to examine the difference between the two different means for fragmenting carbohydrates. Dotted lines are for illustration purposes of the fragmentation peaks.

Imaging Ion Mobility-Mass Spectrometry

377

2. Materials 1. Sample prepared for MALDI analysis (thin tissue section washed, fixed to MALDI plate, and MALDI matrix applied, see Chapters 4, 7, 11, 16, 20, and 21 for detailed methodologies). 2. Mass and drift tube IM standards/calibrants. Mass standards correspond to peptides and proteins bracketing the mass range of interest. Ion mobility structural standards for DTIM are typically C60 and C70 fullerenes, because they exist in one structural form. These can be used for evaluating DTIM resolution and for day-to-day evaluation of instrument performance. Additionally, fullerenes can be used as mass standards as they are structurally separated from biomolecules in conformation space (see Fig. 21.4) and provide a wide range of gas-phase reaction products resulting in peaks spanning a large mass range in increments of 24 Da. To validate gas pressure in DTIM, typically the peptide bradykinin (RPPGFSPFR) is used to compare collision cross-section measurement with the accepted value of 242±2 Å2 (12). Bradykinin can be mixed with matrix of choice or a 1 mg/ml standard solution in H2 O can be combined 1:1 v/v with 20 mg/ml α-cyano-4-hydroxycinnamic acid in 50% methanol. Both calibrants can be applied to MALDI plate using the dried droplet method (55). 3. Traveling wave IM standards/calibrants. As discussed in Section 1.3, estimated collision cross sections obtained by TWIM require internal standards with corresponding absolute collision cross-section values obtained using DTIM. Published absolute collision cross sections can be obtained from several published databases, including (i) peptide collision cross sections determined by ESI (56, 57), (ii) intact protein collision cross sections determined by ESI (58), (iii) peptide collision cross sections determined by MALDI (36), and (iv) biologically relevant carbohydrate, lipid, and oligonucleotide collision cross sections determined by MALDI (59).

3. Methods 3.1. Performing Collision Cross-Section Measurements Using DTIM

1. In order to take measurements, the samples for imaging (tissue, etc.) should be prepared the same as for conventional imaging MALDI-MS (see Note 5). 2. Following insertion of the sample target into the instrument, mass and ion mobility standard/calibrants are measured. In

378

McLean, Fenn, and Enders

particular, to MALDI-IM-MS methods the laser pulse serves as the start signal (t0 ) for measuring the IM arrival time distribution (tatd ). These time distinctions are necessary for the calculations in Step 4. 3. Following separation in the IM drift cell filled with an inert gas (1–10 torr, see Note 6), ions are directed through a skimming and differential pumping region where the pressure is reduced from 1–10 to ∼10−8 torr for mass analysis in the orthogonal TOFMS. The stop time for tatd corresponds to the ion injection time for the TOFMS measurement. 4. To perform the calculations as described in Section 1.2.1 (e.g., equation [4]) the arrival time distribution must be corrected for time spent in regions outside of the drift cell (i.e., time spent traversing from the MALDI plate into the drift cell, in skimming and differential pumping regions, and ion optic regions prior to the source of the TOFMS). This will result in the drift time (td ) of the ions within the IM drift cell used in the calculation of collision cross section:

td = tatd − tdtc 5. To determine the value of tdtc , IM separations are performed by varying the voltage across the drift cell while maintaining all other experimental parameters constant. The arrival time distribution measured at each drift voltage is then plotted versus the inverse of drift voltage (1/V). Provided the range of voltages used maintains ion separations under lowfield conditions, this plot will result in a linear correlation. If non-linearity is observed, a calculation of the low-field limit should be performed (see Note 1), because curvature in this plot indicates that mobility is not constant over the voltage range used. A linear regression of this data results in a y-intercept corresponding to tdtc (see Note 7). Preferably at least five voltages should be used to define this line although for high-precision measurements as many voltages as is practical should be used (see Note 8). 6. After the td has been determined from the tatd , it can now be used to calculate the collision cross section, , of the ion of interest through the equation [5] (see Notes 9 and 10 (26)). 7. After the collision cross section has been calculated, this can be further related to the structure using molecular dynamic simulations. More information about these computational methods can be found in more detail in other resources (31–34). An excellent overview and tutorial of these strategies can be found elsewhere (60).

Imaging Ion Mobility-Mass Spectrometry

379

8. For calculating relative collision cross sections using traveling wave ion mobility-MS, there are two main procedures used which can be found in the literature (29, 30).

4. Notes 1. For electrostatic fields higher than the low-field limit, the ion velocity distribution depends less strongly on the temperature of the separation and the mean ion energy increases as it traverses the drift region. Consequently, K is no longer constant and depends on the specific ratio of the electrostatic field to the gas number density (E/N) (see (26) for a derivation of calculating the low-field limit for a particular analyte). 2. In principle, the Synapt has sufficient activation/ dissociation regions to perform up to MS5 , although typically up to MS3 is practically feasible. 3. Presentation of 3D conformation space data (IM arrival time distribution, m/z, signal intensity) is typically projected with false coloring or gray scale representing signal intensity to project 3D data in a 2D plot. 4. There is presently no consensus on the reporting of IM-MS conformation space data, i.e., TWIM-MS data are generated with arrival time distribution on the abscissa and m/z on the ordinate axes, but it is either reported using this convention or where the axes are inverted. This reporting of conformation space data parallels historical preferences in the reporting of DTIM-MS data (e.g., see Figs. 21.3, 21.4, and 21.5). 5. Typically ionization is performed at the pressure of the DTIM (e.g., 1–10 torr), which results in moderate pressure MALDI. Thus some collisional cooling typically takes place after ionization and can result in matrix-adducted and cluster species. These can be dissociated prior to DTIM by performing injected ion experiments. Furthermore, matrix optimization may be required. One effect of moderate pressure MALDI that we have observed is that higher ion currents can be achieved at slightly lower matrix-to-analyte ratios (i.e., 1,000–100:1) than those used in high-vacuum MALDI (i.e., 10,000–1,000:1). 6. As developed in more detail in the theory section, typically He is used because of its low mass and low polarizability relative to other inert gases. However, other drift gases or

380

McLean, Fenn, and Enders

drift gas additives can be used to promote long-range interactions between the ion and drift gas. This is analogous to tuning selectivity in HPLC by changing the mobile or stationary phase that is used. 7. The y-intercept of this plot corresponds to tdtc because it represents the limit of td →0 at infinite drift cell voltage. Also note that the accuracy with which this correction should be made is more important for shorter drift times and its significance is less important at longer drift times. For example, in fast separations as described for imaging IM-MS experiments the values of tdtc can approach the relative magnitude of td . 8. For the most accurate results, the drift time correction should be evaluated for each component in the IM profile. The motivation for evaluating individual drift time corrections arises from additional ion–neutral collisions in the differential pumping regions at the entrance and/or exit of the IM drift cell. In these regions the gas dynamics typically transition from viscous to molecular flow, e.g., at the exit aperture of the drift cell at 1–10 torr to the high vacuum (∼10−8 torr) of the mass spectrometer, respectively. 9. Note that the equation for calculating collision cross section is derived from classical electrodynamics, and as such, great care should be exercised in the dimensionality of the units used. Specifically, the units for E should be expressed in CGS Gaussian units, i.e., statvolts cm−1 , where 1 statvolt equals 299.79 V. Note that statvolts cm−1 is equivalent to statcoulombs cm−2 and that elementary charge, e, is 4.80×10−10 statcoulombs. 10. By comparing empirically determined cross sections with theoretical results, it has been shown that the hard-sphere approximation is best suited for analytes larger than ca. 1,000 Da, which is typically the size range in which many biological measurements are made. However, as the size of the analyte approaches the size scale of the drift gases used for separation, long-range interaction potential between the ion and neutral must be considered for accurate results (33, 61, 62).

Acknowledgments We thank Whitney B. Ridenour and Richard M. Caprioli (Vanderbilt University) for assistance and use of the Synapt HDMS (data shown in Figs. 21.6 and 21.7), which is supported

Imaging Ion Mobility-Mass Spectrometry

381

by the Vanderbilt University Mass Spectrometry Research Core. Financial support for this work was provided by the National Institutes of Health-NIDA (#HHSN271200700012C), Vanderbilt University College of Arts and Sciences, Vanderbilt Institute of Chemical Biology, Vanderbilt Institute for Integrated Biosystems Research and Education, the American Society for Mass Spectrometry (Research award to J.A.M), the Spectroscopy Society of Pittsburgh, Waters Corp., and Ionwerks Inc. References 1. Fenn, L. S., McLean, J. A. (2008) Biomolecular structural separations by ion mobilitymass spectrometry. Anal Bioanal Chem, 391, 905–909. 2. McLean, J. A., Ruotolo, B. T., Gillig, K. J., Russell, D. H. (2005) Ion mobilitymass spectrometry: a new paradigm for proteomics. Int J Mass Spectrom, 240, 301–315. 3. Wyttenbach, T., Bowers, M. T. (2003) Gasphase conformations: the ion mobility/ion chromatography method. Modern Mass Spectrom, 225, 207–232. 4. Jarrold, M. F. (2000) Peptides and proteins in the vapor phase, Annu Rev Phys Chem, 51, 179–207. 5. Clemmer, D. E. Jarrold, M. Fd. (1997) Ion mobility measurements and their applications to clusters and biomolecules. J Mass Spectrom, 32, 577–592. 6. Eiceman, E. A., Karpas, Z. (2004) Ion Mobility Spectrometry, 2nd Ed., CRC Press, Boca Raton, FL, Chapter 1. 7. McAfee, K. B., Jr., Edelson, D. (1963) Identification and mobility of ions in a townsend discharge by time-resolved mass spectrometry, Proc Phys Soc Lond, 81, 382–384. 8. Barnes, W. S., Martin, D. W., McDaniel, E. W. (1961) Mass spectrographic identification of the ion observed in hydrogen mobility experiments, Phys Rev Lett, 6, 110–111. 9. Gieniec, J., Mack, L. L., Nakamae, K., Gupta, C., Kumar, V., Dole, M. (1984) Electrospray mass spectroscopy of macromolecules: application of an ion-drift spectrometer. Biomed Mass Spectrom, 11, 259–268. 10. Von Helden, G., Wyttenbach, T., Bowers, M. T. (1995) Inclusion of a MALDI ion source in the ion chromatography technique: conformation information on polymer and biomolecular ions. Int J Mass Spectrom Ion Process, 146/147, 349–364. 11. Shelimov, K. B., Clemmer, D. E., Hudgins, R. R., Jarrold, M. F. (1997) Protein structure in vacuo: the gas phase conformations

12.

13.

14.

15.

16.

17.

18.

of BPTI and cytochrome c, J Am Chem Soc, 119, 2240–2248. Wyttenbach, T., Von Helden, G., Bowers, M. T. (1996) Gas-phase conformation of biological molecules: bradykinin. J Am Chem Soc, 118, 8355–8364. Von Helden, G., Wyttenbach, T., Bowers, M. T. (1995) Conformation of macromolecules in the gas phase: use of matrix-assisted laser desorption methods in ion chromatography. Science, 267, 1483–1485. Myung, S., Lee, Y. J., Moon, M. H., Taraszka, J., Sowell, R., Koeniger, S., Hilderbrand, A. E., Valentine, S. J., Cherbas, L., Cherbas, P., Kaufmann, T. C., Miller, D. F., Mechref, Y., Novotny, M. V., Ewing, M. A., Sporleder C. R., Clemmer, D. E. (2003) Development of high-sensitivity ion trap ion mobility spectrometry time-of-flight techniques: a high-throughput nano-LC-IMSTOF separation of peptides arising from a Drosophila protein extract. Anal Chem, 75, 5137–5145. Isailovic, D., Kurulugama, R. T., Plasencia, M. D., Stokes, S. T., Kyselova, Z., Goldman, R., Mechref, Y., Novotny, M. V. and Clemmer, D. E. (2008) Profiling of human serum glycans associated with liver cancer and cirrhosis by IMS-MS. J Proteome Res, 7, 1109–1117. Liu, X., Valentine, S. J., Plasencia, M. D., Trimpin, S., Naylor, S., Clemmer, D. E. (2007) Mapping the human plasma proteome by SCX-LC-IMS-MS. J Am Soc Mass Spectrom, 18, 1249–1264. Valentine, S. J., Plasencia, M. D., Liu, X., Krishnan, M., Naylor, S., Udseth, H. R., Smith, R. D., Clemmer, D. E. (2006) Toward plasma proteome profiling with ion mobility-mass spectrometry. J Proteome Res, 5, 2977–2984. Liu, X., Plasencia, M., Ragg, S., Valentine, S. J., Clemmer, D. E. (2004) Development of high throughput dispersive LC–ion

382

19.

20.

21.

22.

23.

24.

25.

26. 27.

28.

29.

30.

McLean, Fenn, and Enders mobility–TOFMS techniques for analysing the human plasma proteome. Brief Funct Genomic Proteomic, 3, 177–186. Liu, X., Miller, B. R., Rebec, G. V., Clemmer, D. E. (2007) Protein expression in the striatum and cortex regions of the brain for a mouse model of Huntington’s disease.J Proteome Res, 6, 3134–3142. Taraszka, J. A., Kurulugama, R., Sowell, R. A., Valentine, S. J., Koeniger, S. L., Arnold, R. J., Miller, D. F., Kaufman, T. C., Clemmer, D. E. (2005) Mapping the proteome of Drosophila melanogaster: analysis of embryos and adult heads by LC-IMS-MS methods. J Proteome Res, 4, 1223–1237. Benesch, J. L. P., Ruotolo, B. T., Simmons, D. A., Robinson, C. V. (2007) Protein complexes in the gas phase: technology for the structural genomics and proteomics. Chem Rev, 107, 3544–3567. Ruotolo, B. T., Hyung, S.-J., Robinson, P. M., Giles, K., Bateman, R. H., Robinson, C. V. (2007) Ion mobility-mass spectrometry reveals long-lived, unfolded intermediates in the dissociation of protein complexes. Angew Chem Int Ed, 46, 8001–8004. Ruotolo, B. T., Giles, K., Campuzano, I., Sandercock, A. M., Bateman, R. H., Robinson, C. V. (2005) Evidence for macromolecular rings in the absence of bulk water. Science, 310, 1658–1661. Jackson, S. N., Ugarov, M., Egan, T., Post, J. D., Langlais, D., Schultz, J. A., Woods, A. S. (2007) MALDI-ion mobility-TOFMS imaging of lipids in rat brain tissue. J Mass Spectrom, 42, 1093–1098. McLean, J. A., Ridenour, W. B., Caprioli, R. M. (2007) Profiling and imaging of tissues by imaging ion mobility-mass spectrometry. J Mass Spectrom, 42, 1099–1105. Mason, E. A., McDaniel, E. W. (1988) Transport Properties of Ions in Gases. John Wiley & Sons, New York, NY. Mason, E. A. (1984) Ion mobility: its role in plasma chromatography, in Plasma Chromatography, (Carr, T. W. ed.), Plenum Press, New York, NY, 43–93. Revercomb, H. E. and Mason, E. A. (1975) Theory of plasma chromatography/gaseous electrophoresis – a review. Anal Chem, 47, 970–983. Ruotolo, B. T., Benesch, J. L. P., Sandercock, A. M., Hyung, S.-J. Robinson, C. V. (2008) Ion mobility-mass spectrometry analysis of large protein complexes. Nat Protoc, 3, 1139–1152. Williams, J. P., Scrivens, J. H. (2008) Coupling desorption electrospray ionization

31.

32.

33.

34.

35.

36.

37.

38.

39. 40.

41. 42.

and neutral desorption/extractive electrospray ionization with a travelling-wave based ion mobility mass spectrometer for the analysis of drugs. Rapid Commun Mass Spectrom, 22, 187–196. Gidden, J., Bowers, M. T. (2003) Gas-phase conformations of deprotonated and protonated mononucleotides determined by ion mobility and theoretical modeling. J Phys Chem B, 107, 12829–12837. Gidden, J., Bowers, M. T. (2003) Gasphase conformations of deprotonated trinucleotides (dGTT(–), dTGT(–), and dTTG (–)): the question of zwitterion formation. J Am Soc Mass Spectrom, 14, 161–170. Wyttenbach, T., Witt, M., Bowers, M. T. (2000) On the stability of amino acid zwitterions in the gas phase: the influence of derivatization, proton affinity, and alkali ion addition. J Am Chem Soc, 122, 3458–3464. Shvartsburg, A. A., Jarrold, M. F. (1996) An exact hard-spheres scattering model for the mobilities of polyatomic ions. Chem Phys Lett, 261, 86–91. Dwivedi, P., Wu, P., Klopsch, S. J., Puzon, G. J., Xun, L., Hill, H. H. (2008) Metabolic profiling by ion mobility mass spectrometry (IMMS). Metabolomics, 4, 63–80. Tao, L., McLean, J. R., McLean, J. A., Russell, D. H. (2007) A collision crosssection database of singly-charged peptide ions. J Am Soc Mass Spectrom, 18, 1232– 1238. Ruotolo, B. T., Verbeck, G. F., Thomson, L. M., Woods, A. S., Gillig, K. J., Russell, D. H. (2002) Distinguishing between phosphorylated and nonphosphorylated peptides with ion mobility-mass spectrometry. J Proteome Res, 1, 303–306. Furche, F., Ahlrichs, R., Weis, P., Jacob, C., Gilb, S., Bierweiler, T., Kappes, M. M. (2002) The structures of small gold cluster anions as determined by a combination of ion mobility measurements and density functional calculations. J Chem Phys, 117, 6982–6990. Shvartsburg, A. A., Smith, R. D. (2008) Fundamentals of traveling wave ion mobility spectrometry. Anal Chem, 80, 9689–9699. Mason, E. A., McDaniel, E. W. (1988) Measurement of drift velocities and longitudinal diffusion coefficients, in Transport Properties of Ions in Gases, John Wiley & Sons, New York, NY, 31–102. Kanu, A. B., Dwivedi, P., Tam, M., Matz, L., Hill, H. H., Jr. (2008) Ion mobility-mass spectrometry, J Mass Spectrom, 43, 1–22. Steiner, W. E., Clowers, B. H., English, W. A., Hill, H. H., Jr. (2004) Atmospheric

Imaging Ion Mobility-Mass Spectrometry

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

pressure matrix-assisted laser desorption/ionization with analysis by ion mobilitymass spectrometry. Rapid Commun Mass Spectrom, 18, 882–888. McLean, J. A., Russell, D. H. (2003) Subfemtomole peptide detection in ion-mobilitytime-of-flight mass spectrometry measurements. J Proteome Res, 2, 427–430. McLean, J. A., Ridenour, W. B., Caprioli, R. M. (2007) Imaging ion mobility mass spectrometry: advantages, challenges, and future prospects. Proceedings of the 55th Annual Meeting of the American Society for Mass Spectrometry, Indianapolis, IN. Merenbloom, S. I., Glaskin, R. S., Clemmer, D. E. (2009) High resolution ion cyclotron mobility spectrometry. Anal Chem, 81, 1482–1487. Wyttenbach, T., Kemper, P. R., Bowers, M. T. (2001) Design of a new electrospray ion mobility mass spectrometer. Int J Mass Spectrom, 212, 13–23. Dugourd, Ph., Hudgins, R. R., Clemmer, D. E., Jarrold, M. F. (1997) High-resolution ion mobility measurements. Rev Sci Instrum, 68, 1122–1129. Giles, K., Pringle, S. D., Worthington, D. L., Wildgoose, J. L., Bateman, R. H. (2004) Applications of travelling wave-based radiofrequency-only stacked ring ion guide. Rapid Commun Mass Spectrom, 18, 2401–2414. Pringle, S. D., Giles, K., Wildgoose, J. L., Williams, J. P., Slade, S. E., Thalassinos, K., Bateman, R. H., Bowers, M. T., Scrivens, J. H. (2007) An investigation of the mobility separation of some peptide and protein ions using a new hybrid quadrupole/travelling wave IMS/oa-ToF instrument. Int J Mass Spectrom, 261, 1–12. Vakhrushev, S. Y., Langridge, J., Campuzano, I., Hughes, C., Peter-Katalinic, J. (2008) Ion mobility mass spectrometry analysis of human glycourinome. Anal Chem, 80, 2506–2513. Riba-Garcia, I., Giles, K., Bateman, R.H., Gaskell, S.J. (2008) Evidence for structural variants of a- and b- type peptide fragment ions using combined ion mobility/mass spectrometry. J Am Soc Mass Spectrom, 19, 609–613. Trim, P. J., Henson, C. M., Avery, J. L., McEwen, A., Snel, M. F., Claude, E., Marshall, P. S., West, A., Princivalle, A. P., Clench, M. R. (2008) Matrix-assisted laser desorption/ionization-ion mobility

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

383

separation-mass spectrometry imaging of vinblastine in whole body tissue sections. Anal Chem, 80, 8628–8634. Reyzer, M. L., Hsieh, Y., Ng, K., Korfmacher, W .A., Caprioli, R. M. (2003) Direct analysis of drug candidates in tissue by matrix-assisted laser desorption/ionization mass spectrometry. J Mass Spectrom, 38, 1081–1092. Domon, B., Costello, C. E. (1988) A systematic nomenclature for carbohydrate fragmentations in FAB-MS/MS spectra of glycoconjugates. Glycoconjugate J, 5, 397–409. Karas, M., Hillenkamp, F. (1988) Laser desorption ionization of proteins with molecular masses exceeding 10,000 daltons. Anal Chem, 60, 2299–2301. Henderson, S. C., Valentine, S. J., Counterman, A. E., Clemmer, D. E. (1999) ESI/ion trap/ion mobility/timeof-flight mass spectrometry for rapid and sensitive analysis of biomolecular mixtures. Anal Chem, 71, 291–301. Hoaglund, C. S., Valentine, S. J., Sporleder, C. R., Reilly, J. P., Clemmer, D. E. (1998) Three-dimensional ion mobility/TOFMS analysis of electrosprayed biomolecules. Anal Chem, 70, 2236–2242. Clemmer cross section database (2006) (Assessed January 28th, 2009 at http:// www.indiana.edu/∼clemmer/Research/cross section database/cs database.htm) Fenn, L. S., Kliman, M., Mahsutt, A., Zhao, S. R., McLean, J. A. (2009) Characterizing ion mobility-mass spectrometry conformation space for the analysis of complex biological samples. Anal Bioanal Chem, 394, 235–244. Theories and analysis of IM-MS data (Assessed January 28th, 2009 at http:// bowers.chem.ucsb.edu/theory_analysis/index. shtml). Wyttenbach, T., von Helden, G., Batka, J. J., Jr., Carlat, D., Bowers, M. T. (1997) Effect of the long-range interaction potential on ion mobility measurements. J Am Soc Mass Spectrom, 8, 275–282. Mesleh, M. F., Hunter, J. M., Shvartsburg, A. A., Schatz, G. C., Jarrold, M. F. (1996) Structural information from ion mobility measurements: effects of the long-range potential. J Phys Chem, 100, 16082–16086.

Chapter 22 Tutorial: Multivariate Statistical Treatment of Imaging Data for Clinical Biomarker Discovery Sören-Oliver Deininger, Michael Becker, and Detlev Suckau Abstract Cancer research is one of the most promising application areas for the new technology of MALDI tissue imaging. Cancerous tissue can easily be distinguished from healthy tissue by dramatically changed metabolism, growth, and apoptotic processes. Of even higher interest is the fact that MALDI imaging allows to unveil molecular differentiation undetectable by classical histological techniques. Thus, MALDI imaging has tremendous potential as a tool to characterize the therapeutic susceptibility of tumors in biopsies as well as to predict tumor progression in endpoint studies. However, some aspects are important to consider for a successful MALDI imaging-based cancer research. Cancer sections are usually very heterogeneous – different biochemical pathways can be active in individual tumor clones, at different development stages or in various tumor microenvironments. Understanding tissue at this level is only possible for experienced histopathologists working on high-resolution optical images. Therefore, the largest benefit from the use of MALDI imaging results in histopathology will arise if molecular images are related to classical high-resolution histological images in a simple way without the need to interpret mass spectra directly. Each MALDI imaging data set effectively provides information on hundreds of molecules and permits the generation of molecular images displaying the relative abundance of these molecules across the tissue. The interpretation of these in the histological context is a major challenge in terms of expert analysis time. This is true especially for clinical work with hundreds of tissue specimens to be analyzed by MALDI, interpreted, and compared. Therefore, a MALDI imaging workflow is described here that enables fast and unambiguous interpretation of the MALDI imaging data in the histological context. Preprocessing of the image data using statistical tools allows efficient and straightforward interpretation by the histopathologist. In this chapter, we explain the use of principal component analysis (PCA) and hierarchical clustering (HC) for the efficient interpretation of MALDI imaging data. We also outline how these methods can be used to compare specific disease states between patients in the search for biomarkers. Key words: MALDI imaging, molecular histology, tumor, principal component analysis, hierarchical clustering, biomarker.

S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_22, © Springer Science+Business Media, LLC 2010

385

386

Deininger, Becker, and Suckau

1. Introduction Cancer research is one of the most promising application areas for the new technology of MALDI tissue imaging (for review, see, e.g., (1)). Cancerous tissue can easily be distinguished from healthy tissue due to dramatically changed metabolism, growth, and apoptotic processes. In addition, tumors themselves are highly heterogeneous and can consist of areas of invasive tumor, the invasion front, tumor cells of different development stages, lymphocyte infiltrations, inflamed tissue, and others. Different metabolic states inside a tumor, such as hypoxic areas, or differentially regulated cellular proteases further complicate the overall picture. Non-tumorous tissue in the vicinity of the tumor cells, the stroma, may influence the tumor as well. MALDI imaging is able to unveil molecular differentiation undetectable by classical histological techniques and can therefore serve as a tool to characterize tumor tissue from biopsies. In differential studies, therapeutic susceptibility of tumors as well as tumor progression in endpoint studies can be assessed (2). Understanding MALDI imaging results at this level of heterogeneity is a task for experienced pathologists and usually requires the parallel evaluation of histologically stained tissue at microscopic resolution. A major challenge for MALDI imaging in clinical research is the correlation of histology and the molecular information, i.e., the precise matching of two image dimensions. A further challenge is the efficient interpretation of MALDI imaging data in the histological context, which is of particular importance in clinical studies where large numbers of specimen are analyzed. Peak-by-peak interpretation of MALDI imaging data is time-consuming and tedious. In fact, the detailed analysis of numerous tissue sections by a pathologist, as necessary in clinical studies, is not feasible due to time constraints. One way to eliminate this limitation in inspection time has been suggested by (3). In this workflow, the pathologist defines specific tissue regions based on conventional histology prior to the MALDI imaging analysis. Using these assignments, it is then possible to compare patient samples to look for disease (stage) markers. However, a major advantage of MALDI imaging is the ability to detect molecular differentiation not apparent from histology alone. Although of particular interest, this information is not fully exploited when the interpretation of the molecular images is left in the hands of a mass spectrometrist. It is mandatory that both the histological and the molecular information in MALDI imaging experiments are evaluated by the pathologists themselves. Such evaluation on a reasonable timescale requires the use of statistical methods that reduce the

Tutorial: Multivariate Statistical Treatment of Imaging Data

387

molecular data in a comprehensive way. One commonly used way to achieve this is the reconstruction of images based on principal component analysis (PCA). This approach has long been used in secondary ion mass spectrometry (4–6) and is gaining popularity in MALDI imaging (7, 8). Using PCA, the information of a MALDI imaging data set is condensed to a small number of molecular images allowing automated feature extraction. It is a very robust unsupervised multivariate approach to report the results of MALDI imaging experiments if no further information on the tissue is available. However, PCA does not increase the detail level of the analysis or enable interactive work with the data set. These limitations are overcome by hierarchical clustering combined with a user interface that allows the reconstruction of images based on spectra similarity (1, 7), a method which allows to interactively explore the data set in an efficient way based on both the histological and the molecular information. In a clinical study, this allows the pathologist to select the mass spectra that are characteristic for a certain disease state, which in a later step can be compared across different patients. In this chapter, we explain the use of principal component analysis (PCA) and hierarchical clustering (HC) for the efficient interpretation of MALDI imaging data (7, 9). We also outline how these methods can be used to compare specific disease states between patients in the search for biomarkers. 1.1. Molecular Histology

The initial interpretation of a tumor section is always based on histology, with hematoxylin and eosin (H&E) staining being the gold standard. Recognizing the state of a given cell or cell population requires microscopical evaluation at a magnification that allows to see the shape of the nuclei. Unstained tissue sections are not helpful in this regard as they typically do not display sufficient contrast, so histological staining is mandatory to assign tissue features. The result of a MALDI imaging experiment can be interpreted only in conjunction with the histology. Consequently, a major hurdle for MALDI imaging in cancer research is the ability for “multimodal imaging,” i.e., the capability to derive different types of image information (H&E stain and MALDI) from the same tissue section. Therefore, a major breakthrough toward clinical MALDI imaging was the discovery that the MALDI matrix can be washed off after an experiment and histologically stained images can then be obtained from the same section (1, 10, 11). For practical use, the sample has to be prepared on a slide that addresses the requirements for both imaging methods: (1) optical transparency for transmission microscopy and (2) electrical conductivity for MALDI analysis. Conductive indium-tin-oxide (ITO)coated glass slides fulfill both requirements and have become the de facto standard.

388

Deininger, Becker, and Suckau

As classical histological images can be directly overlaid with the molecular dimension derived from the same tissue section, this type of experiment can truly be called molecular histology. Earlier reports demonstrated that some histological stains were compatible with subsequent MALDI imaging analyses (12). Amongst them were methylene blue, cresyl violet but not H&E. Unfortunately, the MALDI-compatible stains do not yield the same information content as the H&E staining. In such prestaining approaches, unambiguous assignment of MALDI imaging results to tissue features is possible, but to understand the nature of the observed features, an H&E staining of a serial section is still necessary. The use of consecutive sections for MALDI imaging and the H&E staining has also been reported (3). The section used for MALDI imaging is scanned prior to the matrix application and this image is overlaid with a H&E stain of a consecutive section. Apart from the fact that small features such as carcinoma in situ can be present in one section but absent or significantly shifted in the other section, obtaining consecutive sections of good quality (e.g., not distorted, ripped, or folded) represents a technical challenge, as even slight differences can make the overlay difficult. In conclusion, the “post-MALDI staining” workflow with MALDI imaging as the first and removal of the matrix, H&E staining, and microscopy as the second step is now the established and preferred approach to MALDI imaging in the clinical context, because it compromises neither the molecular nor the histological dimension of the analysis. 1.2. Interpreting MALDI Images by Principal Component Analysis (PCA)

MALDI imaging data are multidimensional data, with each mass signal defining one molecular dimension. The aim of PCA is to reduce the dimensionality of the molecular information to a small number of relevant dimensions. (For a detailed explanation of PCA, see Section 1.5) Such PCA results can be turned into image information by displaying the scores of selected principal components for each pixel (i.e., each spectrum) as color density plots. As a result, a rather small number of images (i.e., one per principal component observed) are obtained which combine the relevant information, rather than hundreds of single-mass ion images (see Figs. 22.1a and 22.2a). The main advantage of PCA is that it allows unsupervised feature extraction from MALDI imaging data without any understanding of the tissue. The image based on the first principal component (PC1) will show the main variance in the data set and can be expected to be in good agreement with histology. This is true if the MALDI imaging analysis is performed only within the boundaries of the tissue or if only mass spectra from the tissue region are subjected to the PCA calculation. If the imaging software allows acquisition of spectra from a rectangular region only, the main

Tutorial: Multivariate Statistical Treatment of Imaging Data

389

Fig. 22.1. PCA and hierarchical clustering example shown on a kidney sample. (a) Optical image of the kidney section prior to matrix application. (b) Scores of first principal component. The image is in agreement with the anatomy of the kidney, with the renal pelvis and part of the cortex showing the hot colors. (c) Clustering result: Renal cortex, medulla, and pelvis are defined by the highest level clusters. The advantage of this type of analysis is that the dendrogram nodes can be expanded and highlighted until the desired molecular structure is found. Scale bar: 2 mm.

variance in the data comes from “on-tissue” versus “off-tissue” spectra, so PC1 will mainly reflect this difference. In this case the histology of the section will be better represented by higher order principal components. In a clinical research setting, the person operating the mass spectrometer is typically not able to interpret the results in histological context. Yet the operator of the mass spectrometer may be asked to present initial results of the experiments to the pathologist for a more detailed evaluation. This is where PCA is most helpful. Without understanding of the tissue architecture, one can create a small number of principal component heatmap images that are in good agreement with the histology. Based on these images, it is possible to quickly conclude whether the experiment was successful or if preparation artifacts are present. Further interpretation of imaging data based on PCA is limited. PCA is not able to classify spectra as similar, and moreover spectra that are similar in one principal component may be very different in another. To determine groups of “similar” spectra, other classification or clustering techniques have to be applied, such as the hierarchical clustering discussed here. 1.3. Interpreting MALDI Images by Hierarchical Clustering (HC)

In hierarchical clustering, mass spectra are grouped by similarity in a dendrogram. The distance of the spectra in the multidimensional space is reflected by the distance and position within the hierarchical tree or dendrogram. Moving from the root toward the branches of the hierarchical tree, mass spectra of increasing

390

Deininger, Becker, and Suckau

similarity are grouped into nodes of lower order. MALDI images can now be reconstructed by selecting all mass spectra (i.e., image pixels) from a given node in the dendrogram and assigning a common color to them. The main structure in a data set can therefore be expected in the top nodes near the root of the dendrogram. Several features make the hierarchical clustering especially useful for the interpretation of MALDI imaging data: • The “main” structure of the data set is often found at the top levels. For example, in a tumor data set, top nodes in the dendrogram will most likely separate tumor versus nontumor (see Fig. 22.2c). • A hierarchy of the clusters is maintained, e.g., information on differentiation states of the tumor which are more closely related to each other rather than to non-tumor area in the same tissue is conserved in the result (see Fig. 22.2d).

Fig. 22.2. PCA and hierarchical clustering for a gastric cancer section. (a) H&E-stained tissue section after MALDI imaging measurement. (b) Scores of the first principal component show the hot colors in the tumor area. (c) Hierarchical clustering: Top dendrogram nodes differentiate tumor (green and magenta) versus non-tumor (blue, squamous epithelium in red). (d) The dendrogram can be expanded down the tumor node to evaluate the molecular differentiation inside the tumor. This can also be directly correlated with the histology. This workflow enables the fast and concise selection of mass spectra representative for specific tissue states. Scale bar: 2 mm.

Tutorial: Multivariate Statistical Treatment of Imaging Data

391

Interactive exploration of the molecular results is possible in a quick and effective way. For example, if a dendrogram node is found that contains the “tumor” spectra, these can quickly be further differentiated by looking at the match between lower nodes and more detailed histological features. In summary, MALDI imaging combined with HC has been proven as an efficient way to quickly interpret the molecular information in the histological context. 1.4. Interpretation of Clinical Data

The primary goal of MALDI imaging in clinical research is to put molecular information in the tissue into context with clinical data from numerous patients. Such data include information on survival, outcome, and response to treatment for each patient. Large numbers of analyses have to be performed and evaluated, and the generation of MALDI images is not the endpoint of the analysis, but a tool to assign tissue locations that are specific for a certain tissue (health) state. The comparison of large numbers of samples cannot be done on the image level, but must be performed on selected spectra that are representative for specific states (e.g., invasive tumor, tumor development, or normal epithelium), and MALDI imaging is needed to select these spectra. In these comparisons, PCA and HC can now be used to draw comparisons between patients (see Fig. 22.3). One can also screen for molecules that are representative for certain health states, which may be biomarker candidates. One of the true advantages of the imaging approach is that once such potential biomarkers are found, it is possible to return to the original images and evaluate the location of the markers in the histological context.

1.5. Understanding PCA and HC

In this section we give a practical, non-mathematical approach to the understanding of PCA and HC analysis. The most crucial point for the understanding of either technique is how mass spectra can be represented as multidimensional coordinate systems.

1.5.1. From Mass Spectra to Multidimensional Coordinate Systems

A mass spectrum consists of a given number of peaks (i.e., mass signals). If the spectrum consists of n peaks, then we can create an n-dimensional coordinate system with each axis representing one peak (i.e., one particular mass). We can now plot the intensity (or area) of each peak onto the corresponding axis. This step translated our mass spectrum into one data point in an n-dimensional space (see Fig. 22.4). If we repeat this with a large number of mass spectra, they create a “cloud” in that n-dimensional space (see Fig. 22.5a). The most important part of this concept is that mass spectra that are “similar” are located close together in that n-dimensional space.

392

Deininger, Becker, and Suckau

Tumor

A

Mucosa

B

C

Patient #8 Patient #7 Patient #3 Patient #2 Patient #1 Patient #4 Patient #5 Patient #10 Patient #9 Patient #6

Fig. 22.3. Comparison of imaging data across patients. (a) Pseudo-gel view of selected mass spectra from different images/samples. The mass spectra are representative for tumor or tumor-free mucosa. Spectra in between dashed lines are from the same patient, each of these “lanes” represents one patient. Characteristic biomarkers can now be found by visual inspection or by statistical tools such as receiver-operating curves or p-values. (b) PCA applied to the data set. Each element represents one patient; squares indicate tumor and circle tumor-free mucosa. In this score plot of the first three principal components a separation between tumor and tumor-free mucosa is seen. This indicates that classification based on the MALDI imaging result is possible. (c) Hierarchical clustering can be performed, e.g., on tumor spectra from all patients. This allows correlation of patient clusters with clinical meta-information. The data shown here are described in (7).

1.5.2. Principal Component Analysis (PCA)

In mass spectra constituting an imaging data set, there will always be a correlation of peaks, and we can also assume that similar tissue types will generate similar mass spectra. For this reason, the data points in the n-dimensional space (defined by the peaks of the mass spectra) are not randomly distributed but will have an orderly structure. The aim of the PCA is to transform the original coordinate system in a way that better represents this internal structure. This is achieved by shifting and rotating the original coordinate systems (defined by peaks) so that one axis points into the direction of highest variance in the data set (see Fig. 22.5c). The next axis will be rotated to point in the direction of the second highest variance and so on. The axes of this new

Tutorial: Multivariate Statistical Treatment of Imaging Data

393

Fig. 22.4. This figure shows the most important concept for interpretation of mass spectra by multivariate statistics. (a) Here a mass spectrum with n (in this example 3) peaks is symbolized. The mass spectrum can be represented by the intensities of its peaks. (b) We can now create an n-dimensional coordinate system. Each coordinate represents the intensity of one peak from the original mass spectrum. The original mass spectrum is now represented as a point in an n-dimensional space. Note: The same concept applies if there are more than three peaks. The mathematics will stay the same, although visualization in a simple way is no longer possible. If we apply this to a larger number of spectra, then each spectrum will occupy one point in the n-dimensional space, and points that are located closely together are spectra that are very similar.

coordinate system are called principal components. The value a spectrum assumes on a principal component, e.g., PC1, is called “score.” To reconstruct an image, the score of each spectrum is translated either into a color saturation or into a heatmap representation (see Fig. 22.5e). Since the principal components are ordered in descending overall variance, PC1 contains the most important information in a data set. While PCA shows similar spectra in similar colors, it does not provide any information on how many groups of similar spectra are present or which spectra belong together based on similarity. For this information a clustering algorithm is much better suited. The scaling of the variables (in our case peaks) is an additional important concept for PCA. PCA (as well as the HC discussed later) reveals the main variance in the data set. Usually, peaks with a high average intensity across the data set also have a high (absolute) variance when compared to peaks with low intensity. Thus, peak intensity variance must be scaled prior to PCA, otherwise the results will be dominated by the most intensive peaks only. 1.5.3. Hierarchical Clustering (HC)

During the HC analysis clusters (groups of similar spectra) are established based on the distances of the data points in the n-dimensional space as discussed above. HC works agglomeratively and builds the dendrogram bottom-up. The algorithm works as follows: The two data points with the smallest distance are identified. These two data points then form one cluster, and this cluster is treated as a new data point for the remainder of the calculation. Now the two data points with the next smallest distance are located and assigned to a cluster (see Fig. 22.6). This procedure is repeated until all data points have been assigned to

394

Deininger, Becker, and Suckau

A

B

C

D

E

F

Fig. 22.5. In this figure the concept of PCA and its application to imaging is explained. (a) This cloud of points represents a number of mass spectra (see Fig. 22.3), e.g., from one imaging experiment. Note that there is a clear structure in the data. This is always the case as in reality there are always correlations between peaks. (b) The same data set viewed from a different angle. (c) The principal components. In the PCA the original coordinate system of the data set is transformed (shifted and rotated) in a way that represents the internal structure of the data better. The axes of this transformed coordinate system are called “principal components” (PC). The principal components are then ordered by descending variance. The first PC points in the direction of highest variance, the second PC points in the direction of second highest variance, and so on. At this stage the PCA has not altered the data. (d) Data set after PCA from a different angle. It is now clearly visible that in this example the third principal component contributes only very little information to the explanation of the data set. (e) The third PC, which contains only little information, is taken away. By doing so the data set is projected into a lower-dimensional space but the main information is kept. The value that each data point assumes on each principal component is called its “score,” and the resulting representation is called a “scores plot.” With real-life imaging data sets containing more than a hundred peaks, often the first few PCs already contain 70% or more information of the original data set. (f) By selecting one particular principal component and assigning a color saturation to the scores, we can now assign each spectrum a color intensity representing its score for that PC. This can be used to create an image from the principal component analysis results.

clusters. The length of the branches in a dendrogram represents the distance between the two data points (or the two clusters of the lower hierarchy). The result of any HC will always be a full dendrogram containing all data points, and the dendrogram can always be traced down to single spectra. One drawback of this approach is that the algorithm cannot decide how many clusters represent “real” clusters and how many clusters represent just random differences in the data. Mathematical ways to find

Tutorial: Multivariate Statistical Treatment of Imaging Data

395

Fig. 22.6. Hierarchical clustering. (i) These five elements shall be subjected to hierarchical clustering. (ii) In a first step, the two elements with the smallest distance are found. Here these are elements A and C. These two elements are then put together into one cluster. The cluster becomes a new element. (iii) Now the elements with the second smallest distance are found. In this case these are elements B and E. These two elements are put together into the second cluster and this cluster also becomes a new element. (iv) Now the elements with the next smallest distance are found, these are cluster AC and the element D. They are grouped together into one cluster. (v) Now only two clusters are left. These are put together into the top-level cluster that contains all elements. Note: The length of the branches in the dendrogram indicates the distance between the two elements in the respective cluster.

the real clusters exist, such as plotting the variance explained by the clusters versus the number of clusters. Such a plot shows an “elbow” at the point where random fluctuation starts (“elbow criterion”). Other methods include repeatedly re-sampling random sub-samples of the original data to locate invariant clusters (i.e., bootstrapping) (13). For the interpretation of MALDI imaging data this does not matter since here the interpretation can be based on the histological knowledge. The pathologist can simply browse into the dendrogram until the histology is sufficiently explained. Selected nodes in the dendrogram can be further explored if they represent possible additional information that is not found by the histology alone. Then the dendrograms can be annotated with their respective tissue state (such as “invasive tumor,” “carcinoma in situ,” “lymphocytes,”). Based on these annotations, spectra of a tissue type can now be compared across different patients. For a more detailed understanding of the HC it has to be noted that there are different options of parameterization. There are multiple ways in which a distance in a multidimensional space can be calculated, with “Euclidean” being the most intuitive one. Other distance metrics include correlation, cosine, Minkowsky,

396

Deininger, Becker, and Suckau

and city-block. Also, there are different options for calculating the distance between two clusters, the so-called linkage. Here the possibilities include the smallest possible distance between two data points in the two clusters (single linkage), the distance between the centers of gravity, the average of all pair-wise distances between two data points (average linkage). The particularly useful “Ward” linkage method tries to minimize the “within cluster” variance. A detailed discussion of these options is beyond the scope of this chapter. 1.5.4. Relationship of PCA and HC

HC and PCA work on any multidimensional data set. Therefore, clustering can be conducted on the original data set as well as on the data set after the PCA transformation. The latter allows to cluster-analyze a data set that is already reduced in the number of dimensions. We found this useful for the clustering calculation with the Euclidean distance.

2. Materials 2.1. Tissue Preparation

1. Conductive indium-tin-oxide (ITO)-coated glass slides (Bruker). 2. Optimum cutting temperature polymer (OCT, Sakura). 3. Two coplin jars with 70% ethanol (HPLC grade). The first jar should contain fresh solvent. 4. One coplin jar with 96% ethanol (HPLC grade). 5. Liquid paper pen, liquid white-out or similar. 6. A scanner suitable to get an optical image of the unstained tissue section (2400 dpi).

2.2. Matrix Preparation

1. Matrix solution: 10 g/l sinapinic acid (puriss. p.a.; Sigma) in 60% acetonitrile (HPLC grade), 0.2% trifluoroacetic acid, 39.8% water (v/v). It is not necessary to use the ultrapure quality of sinapinic acid. The matrix solution should be prepared weekly and stored in the dark. 2. ImagePrep matrix application device (Bruker) for homogeneous matrix coating of tissue samples with fine crystal morphology (∼20 µm).

2.3. MALDI Measurements

1. MALDI adapter plate (Bruker) to hold the electrically conductive, optically transparent slides in the MALDI ion source. 2. MALDI-TOF or TOF/TOF mass spectrometer (autoflex III or ultraflex III, Bruker) with ScoutMTP ion source and

Tutorial: Multivariate Statistical Treatment of Imaging Data

397

high-resolution precision ion source video camera, 200 Hz SmartBeamTM laser with software adjustable 10–100 µm focus diameter. Equipped with FlexImaging 2.1 software or higher for image data acquisition. 2.4. H&E Staining

1. Coplin jar with 70% ethanol. 2. Eosin Y stock solution 1%: 10 g eosin Y (Sigma), 200 ml water, 800 ml 95% ethanol, stir until eosin is dissolved. Store in the dark at room temperature. 3. Eosin Y working solution 0.25%: 250 ml eosin stock solution, 750 ml 80% ethanol, 5 ml glacial acetic acid. Store at room temperature. 4. Quick hardening mounting medium (e.g., Eukitt, Fluka). 5. Hematoxylin solution according to Meyer (Fluka). 6. Coplin jars with deionized water, 70% ethanol, 80% ethanol, 90% ethanol, two jars with absolute ethanol, xylene (histology grade).

2.5. Bioinformatics

1. FlexImaging software (Bruker) version 2.1 or higher. 2. ClinProTools software (Bruker) version 2.2 or higher.

3. Methods 3.1. Tissue Sections

1. Randomize the order of sample preparation (see Note 1). 2. Put tissue piece onto sample stage, use OCT to glue tissue piece to sample holder. Take care that the OCT does not embed the tissue, in particular it should not touch the cutting plane or blade (see Note 2). 3. Cool down ITO slides in the cryostat (see Note 3). 4. Cut tissue section at 8 µm thickness (see Note 4). 5. Pick up the tissue section with the ITO side of the slide or transfer the tissue section with an artist’s brush onto the slide. 6. Use a finger or the back of the hand to warm the slide from underneath, thus thawing and straightening the tissue section. Keep the section warmed until visibly dry. The slide can now be re-frozen to place more sections or taken out to proceed (see Note 5). 7. Dry the section for 5 min in a desiccator (see Note 6). 8. Fix the tissue section by subsequently washing it for 1 min in 70% ethanol (fresh), again in 70% ethanol, 96% ethanol (see Note 7).

398

Deininger, Becker, and Suckau

9. Dry the section again for 5 min in a desiccator. 10. If the sections have to be shipped or stored for more than 1 week then freeze them at –80◦ C. It is highly recommended not to freeze the sections and proceed immediately. Sections should always be shipped on dry ice (see Note 8). 11. With the liquid paper, mark four teach spots roughly rectangular around the tissue section (see Note 9). 12. Acquire a scan of the unstained tissue section including the teach spots (see Note 10). 3.2. Matrix Application

1. Coat one slide at a time with the ImagePrep device with the standard method settings according to the manufacturer’s instructions (see Notes 11 and 12).

3.3. MALDI Measurement

1. Place the ITO slide with the matrix-coated tissue section into the MALDI adapter plate. 2. Insert the adapter plate into the ion source of the MALDI mass spectrometer. 3. Import the optical image of the unstained tissue section into FlexImaging. 4. Co-register the liquid paper spots between the optical image and the video image of the mass spectrometer. 5. Define the acquisition raster and region for imaging. 6. Start automated image acquisition (see Note 13). 7. Submit ITO slides to standard H&E staining and histology.

3.4. H&E Staining and Histology

1. Choose preferred staining protocol or proceed as follows (see Note 14). 2. Wash off matrix in 70% ethanol (usually 1–2 min) (see Note 15). 3. Stain the tissue section for 5 min in hematoxylin solution. 4. Remove excess staining solution by a dip-wash in deionized water. 5. Rinse the slide in running tap water for 5 min. 6. Wash slide for 1 min in deionized water. 7. Put slide into eosin working solution until the section is sufficiently stained (see Note 16). 8. Wash slide in deionized water. 9. Wash slide for 2 min each in 70% ethanol, 80% ethanol, 90% ethanol, 100% ethanol, again 100% ethanol, xylene (see Note 17). 10. With a glass stir bar put one droplet of mounting medium onto the slide and place coverslip onto tissue.

Tutorial: Multivariate Statistical Treatment of Imaging Data

3.5. PCA and HC on Individual Sections

399

1. Load all spectra from the data set into the ClinProTools software and calculate a PCA and a hierarchical clustering (see Note 18). 2. Scan the stained tissue section. 3. Co-register the image of the stained tissue section. 4. Import PCA and clustering results into FlexImaging (see Note 19). 5. Have a tissue expert interpret the data in the histological context and name the relevant nodes of the dendrogram according to the desired aim of the analysis (see Note 20).

3.6. Search for Biomarkers by Comparing Multiple Patient Sections

1. Decide on the tissue states to be compared (such as “invasive tumor” versus “tumor-free mucosa,” Fig. 22.3). 2. In each individual data set, export a list of spectra that are representative of the desired tissue state(s) as determined in Section 3.5.5. 3. Create a master list containing the spectra from 2 in the desired tissue classes (see Note 21). 4. In ClinProTools activate the option “Support spectra grouping.” This ensures that spectra from one patient are treated as replicates rather than individual samples (see Note 22). 5. Now the available statistical tools (p-values, ROC curves) can be used to search for tissue state-specific biomarker candidates. 6. Have a tissue expert evaluate the masses of the biomarker candidates in the histological context on the individual tissue sections (see Note 23). 7. Optionally use PCA or HC at this level to get an idea about biological variance observed in the data set, on whether or not the tissue classes are separated based on the overall variance in the data set, and to correlate patient clusters with clinical meta-information (see Note 24).

4. Notes 4.1. Tissue Sections

1. Multivariate data analyses are sensitive not only to differences in the investigated groups (such as tumor versus non-tumor in this study) but also to changes in the measurement conditions. Such changes can result from, e.g., aging matrix solution and variations in instrument performance. It is not possible to consider all of these influences

400

Deininger, Becker, and Suckau

in advance. That observed differences in the results are indeed inherent properties of the samples and not impacts of sample preparation or measurement can only be proven if the samples are randomized as soon as possible in the workflow. For example, it is not recommended to first cut all tumor sections and then all non-tumor sections. Likewise, it is never a good idea to conduct the matrix preparation and measurement for all tumor sections first and for all non-tumor sections thereafter. In these examples, it would not be possible to prove that a difference observed between tumor and control samples did not arise from variations during sample handling or measurement. 2. In a collaboration between a mass spectrometry facility and a pathology the tissue sections are usually prepared in the pathology lab. Although the preparation of tissue sections for MALDI imaging is very similar to the procedure for normal histology, there are small but very important changes that can cause trouble. It is of the utmost importance to discuss and sort out these issues before starting any sample preparation. In our experience, it is also best to discuss and explain the requirements of the sample preparation directly with the person actually doing the cutting. OCT is a polymer solution that is widely used to embed tissue specimens for cryosections. It facilitates cutting and is used to fix the tissue to the sample stage of the cryostat. Because of the detrimental effect of OCT to the MALDI imaging results it is important not to use OCT for embedding the tissue, but is equally important to thoroughly clean all surfaces that have been previously contaminated by OCT. OCT can be used to mount the tissue on the sample stage of the cryostat, as long as it is not cut. If the tissue specimens are already embedded in OCT, as much as possible of it should be trimmed off prior to the cutting. 3. It is important to transfer the tissue sections to the frozen glass slide (or to pick it up with the frozen glass slide) (14). This is in contrast to the widely used routine protocol in histology which simply uses warm glass slides. 4. Thicker tissue sections are easier to prepare, but the histological interpretation becomes more difficult. It is also reported that thinner sections improve mass spectra (15). Thinner sections also stick better to the ITO slides during the H&E staining. 5. Starting to warm the slide next to the tissue section and then slowly moving the finger underneath the section ensures that remaining folds in the section are stretched out.

Tutorial: Multivariate Statistical Treatment of Imaging Data

401

6. Vacuum pumps and desiccators are not routine equipment in histology labs, but heating plates to dry the sections are. It is quite tempting in the preparation of the sections to stick to standard histology procedures that dry the tissue sections on a heated plate. In our experience, this has always been detrimental for the MALDI imaging results. In many cases, it will be necessary to bring a vacuum pump and desiccator to the histology lab. 7. We recommend to use new coplin jars and to keep them dedicated to the MALDI imaging, since standard equipment in a histology lab is often contaminated by OCT. If one is not satisfied with the quality of the mass spectra, one can consider to increase the washing times up to 5 min. 8. Whenever possible the tissue sections should be prepared fresh, stored in vacuum at room temperature for only a short time, and measured as soon as possible. They should only be re-frozen after the initial preparation, if they have to be shipped or stored for a longer time. 9. Try to make the edges of the teach marks uneven, this facilitates the teaching for the MALDI measurement later. Water-soluble liquid paper should be preferred since it will not fall off during the H&E staining. 10. Office scanners, preferable with transparency option, are sufficient here. Specialized scanners for photographic slides are best suited. 4.2. Matrix Application

11. The matrix should be applied directly before the measurement. Other matrix application devices can also be used; in this case the matrix solution needs to be one optimized for the device used. Sample preparation can also be done by manually spraying the matrix; however, this method will not be reproducible enough to effectively compare different sections. 12. Sinapinic acid is the matrix of choice for protein measurements. For peptides and lipids it is better to use 2,5-dihydroxybenzoic acid (DHB) or apha-cyano-4hydroxycinnamic acid (HCCA). The latter has the advantage of giving very small crystals, therefore allowing highest lateral resolution.

4.3. MALDI Measurement

13. If multiple sections are available on the sample carrier, a batch acquisition should be considered.

4.4. H&E Staining and Histology

14. The choice of the staining protocol seems to be uncritical, so it is best to use the one routinely used by the pathology lab. The proposed protocol should be considered a

402

Deininger, Becker, and Suckau

suggestion only. Staining times can be considerably different after MALDI imaging as compared to fresh sections. 15. Occasionally, tissue sections may come off the slide during the staining. In this case they are lost. For this reason it is a good idea to have consecutive backup sections available for the staining. In our experience, if the section detaches from the slide, then this happens during the initial washing step to remove the matrix. 16. Getting a feeling for the right staining intensity requires some practice. In this step the section has to be overstained, since in the subsequent washes in deionized water and 70% ethanol, the eosin is partially washed out. 17. The liquid paper teach marks may come off during the xylene wash. They sometimes move on the surface of the slide when the coverslip is put on and may slip onto the tissue. In doubt, remove the teach mark in the xylene bath with a glass bar. Additional teach marks, etched into the glass with a diamond tip pen, can be used to align the stained image later on. These should be placed before scanning the image as well. 4.5. PCA and HC on Individual Sections

18. Ensure that the peak scaling (by the checkbox “normalize peaks”) is switched on. For the clustering, we have the best experience with the settings “Use PCA,” “Reduce dimensions to 70% explained variance,” “Create full tree,” “Distance method: Euclidean,” “Linkage method: Ward.” 19. The display of PCA results usually benefits from the heatmap representation. 20. The cross-fading tool can help to localize the exact position on the H&E stain. To actually evaluate the histology at that position it is necessary to do microscopy on the stained section.

4.6. Search for Biomarkers by Comparing Multiple Patient Sections

21. For the convenient generation of the master list the tool “SIX” (part of ClinProTools 2.2 SR1) is needed. 22. Not marking spectra from a single patient sample as replicates would result in pseudo-replication, which in turn leads to wrong (and way too small) p-values. 23. This possibility to evaluate the potential biomarker masses directly in the histological context is one of the major strengths of MALDI imaging. This allows immediate evaluation whether the mass in question is indeed a tumormarker or inflammation marker. 24. On this level each patient is represented by one (or two) data points.

Tutorial: Multivariate Statistical Treatment of Imaging Data

403

References 1. Walch, A., Rauser, S., Deininger, S.-O., Höfler, H. (2008) MALDI imaging mass spectrometry for direct tissue analysis: a new frontier for molecular histology. Histochem Cell Biol, 130, 421–434. 2. Yanagisawa, K., Shyr, Y., Xu, B. J., Massion, P. P., Larsen, P. H., White, B. C., Roberts, J. R., Edgerton, M., Gonzalez, A., Nadaf, S., Moore, J. H., Caprioli, R. M., Carbone, D. P. (2003) Proteomic patterns of tumour subsets in non-small-cell lung cancer. Lancet, 362, 433–439. 3. Cornett, D. S., Mobley, J. A., Dias, E. C., Andersson, M., Arteaga, C. L., Sanders, M. E., Caprioli, R. M. (2006) A novel histologydirected strategy for MALDI-MS tissue profiling that improves throughput and cellular specificity in human breast cancer. Mol Cell Proteomics, 5, 1975–1983. 4. Aoyagi, S., Kawashima, Y., Kudo, M. (2005) TOF-SIMS imaging technique with information entropy. Nucl Instrum Methods Phys Res Sect B, 232, 146–152. 5. Lockyer, N. P., Vickerman, J. C. (2004) Progress in cellular analysis using ToF-SIMS. Appl Surf Sci, 231, 377–384. 6. Wagner, M. S., Castner, D. G. (2001) Characterization of adsorbed protein films by ToF SIMS with PCA. Langmuir, 17, 4649–4660. 7. Deininger, S.-O., Ebert, M. P., Fütterer, A., Gerhard, M., Röcken, C. (2008) MALDI imaging combined with hierarchical clustering as a new tool for the interpretation of complex human cancers. J Proteome Res, 7, 5230–5236 8. Yao, I., Sugiura, Y., Matsumoto, M., Setou, M. (2008) In situ proteomics with imaging mass spectrometry and principal

9.

10.

11.

12.

13.

14.

15.

component analysis in the Scrapper-knockout mouse brain. Proteomics, 8, 3692–3701. McCombie, G., Staab, D., Stoeckli, M., Knochenmuss, R. (2005) Spatial and spectral correlations in MALDI mass spectrometry images by clustering and multivariate analysis. Anal Chem, 77, 6118–6124. Crecelius, A. C., Cornett, D. S., Caprioli, R. M., Williams, B., Dawant, B. M., Bodenheimer, B. (2005) Three-dimensional visualization of protein expression in mouse brain structures using imaging mass spectrometry. J Am Soc Mass Spectrom, 16, 1093–1099. Schwamborn, K., Krieg, R. C., Reska, M., Jakse, G., Knuechel, R., Wellmann, A. (2007) Identifying prostate carcinoma by MALDI-imaging. Int J Mol Med, 20, 155–159. Chaurand, P., Schwartz, S. A., Billheimer, D., Xu, B. J., Crecelius, A., Caprioli, R. M. (2004) Integrating histology and imaging mass spectrometry. Anal Chem, 76, 1145–1155. Efron, B., Halloran, E., Holmes, S., (1996) Bootstrap confidence levels for phylogenetic trees. Proc Natl Acad Sci U S A, 93, 13429–13434. Schwartz, S. A., Reyzer, M. L., Caprioli, R. M. (2003) Direct tissue analysis using matrix-assisted laser desorption/ionization mass spectrometry: practical aspects of sample preparation. J Mass Spectrom, 38, 699–708. Sugiura, Y., Shimma S., Setou, M. (2006) Thin sectioning improves the peak intensity and signal-to-noise ratio in direct tissue mass spectrometry. J Mass Spectrom Soc Jpn, 54, 45–48.

Chapter 23 Applications of MALDI-MSI to Pharmaceutical Research Brendan Prideaux, Dieter Staab, and Markus Stoeckli Abstract MALDI-MSI has been demonstrated to be a suitable technique in pharmaceutical research for providing information of the distribution of low molecular weight compounds such as drugs and their metabolites within whole-body tissue sections. Important ADME information can be determined by MALDI-MSI analysis of the distribution of drugs and metabolites in whole-body tissue sections taken from animals killed at a range of time points postdose. In this example we applied MALDI-MSI to the localization of a compound and its primary metabolite in whole-body mouse sections. Key words: MALDI-MSI, whole-body tissue imaging, drug and metabolite imaging.

1. Introduction In the drug discovery process essential information can be gleamed from knowing the uptake of a drug to its target as well as its metabolic pathway processes and the sites at which these are occurring within the body following administration. These absorption, distribution, metabolism, and excretion (ADME) data are required to fully understand the efficacy, safety, and thus viability of a compound, and the earlier this is understood during the drug discovery process then increased potential exists for the ability to adapt to potential problems and subsequent time and cost savings. MALDI-mass spectrometric imaging (MALDI-MSI) has been extensively discussed and reviewed in the literature since its introduction over 10 years ago (1) and has now become an established method of localizing a range of analytes in biological S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_23, © Springer Science+Business Media, LLC 2010

405

406

Prideaux, Staab, and Stoeckli

tissues. Examples of analytes which have been studied include the imaging of endogenous and exogenous species such as proteins (2), peptides (3, 4), and lipids (5) as well as low molecular weight xenobiotics including (but not exclusive to) drugs (6–10). Studies have been conducted at high resolution ($100 k for acoustic deposition to a few dollars for a thin-layer chromatography sprayer, and with methods such as electrospray deposition requiring skilled practitioners. Most of these techniques impose a resolution limit of ∼100 µm, which is similar to the resolution limit imposed by the laser focal diameter of commercial mass spectrometers. A second strategy for minimizing the diffusion of analytes during matrix deposition involves fixing the tissue before (7), or during (8), matrix solution deposition. Such solvent fixation of tissues has been standard practice in histology for over 100 years (9–11). Applying a MALDI matrix dissolved in a solution that denatures and precipitates proteins (8), as is presented here in Method 1 (Section 3.1), results in the diffusion of proteins being thermodynamically unfavorable (12). The benefits of combining well-established histology protocols with matrix deposition include minimized disruption of tissue morphology versus common matrix solutions (2) and compatibility with immunohistochemistry methods for validation of mass spectrometry findings. The matrix solution fixation technique presented herein (8) imposes a resolution limit of ∼1 µm, making laser focal diameter the determinant of image resolution. We describe three methods for the preparation of thin tissue sections for in situ MALDI MS imaging of proteins, lipids, and small molecules. Each method can maintain near-cellular resolution during sample preparation. The first method employs a strategy termed “matrix solution fixation,” which simultaneously fixes tissue and deposits MALDI matrix. This method enables rapid preparation and high-resolution (∼30 µm) MALDI imaging of entire thin tissue sections. The second method uses a sensor controlled vibrational vaporization device for layered deposition of 20−100 µm matrix droplets and limits diffusion of soluble analytes to deposited droplet surface area. The third method employs a modified microinjection setup to selectively deposit matrix upon single cells, offering resolution below the laser spot size (10 µm). Specifically, single neurons are labeled with a mixture of MALDI matrix, sinapinic acid, and rhodamine B (for contrast in light microscopy and for visualization with the MALDI microscope). Fluorescence microscopy is used to guide deposition of matrix and rhodamine B upon single fluorescent cells. Notably, matrix microinjection overcomes the resolution limit imposed by the MALDI laser focal diameter. Numerous other methods serve a similar purpose including laser and mechanical

Tissue Preparation for the In Situ MALDI MS Imaging

417

Table 24.1 Spatial resolution of mass spectrometry methods described in this book and in this chapter. References to the chapters and the methods are made in parenthesis

dissection of individual cells (13, 14), ordered stretching of tissues (15), laser oversampling (5), improved lasers and optics (16), and ion microscopy (17–19) (Table 24.1). These and other “tricks of the trade,” including seeding with dry matrix and careful thaw mounting (20), may be considered to augment the methods described herein.

2. Materials 2.1. Matrix Solution Fixation

1. Microm HM525 cryostat (Mikron Instruments Inc., San Marcos, CA, USA). 2. Embedding medium for frozen tissue specimens to ensure optimal cutting temperature (OCT, Sakura Finetek, Torrance, CA, USA). 3. Protein and peptide calibration standards (Bruker Daltonics Billerica, MA, USA). 4. α-Cyano-4-hydroxycinnamic acid (α-CHCA) matrix and sinapinic acid. 5. HPLC grade methanol, water, ethanol, acetonitrile. 6. Trifluoroacetic acid. 7. Indium tin oxide (ITO) coated coverslips with busbars are 1×18 mm with resistivity from 8 to 12 ohms (thickness 1: 0.13–0.17 mm) (SPI Supplies, West Chester, PA, USA).

418

Agar et al.

8. ImmEdge PEN (Vector Laboratories Inc., Burlingame, CA, USA). 9. Custom made coverslip holders. Made from original Bruker’s Microflex stainless steel target, with a machined cavity of 25×25 mm and depth of 300 µm to accommodate glass coverslips of dimensions: 22×22 mm, thickness 1 (0.13–0.17 mm). The coverslips are held on the target by minimal conductive adhesive tape. For users of TOF–TOF instruments such as the Bruker Ultraflex, an ITO-coated glass slide holder and ITO-coated glass slides are available as stock items, and therefore a custom made coverslip holder is unnecessary. 10. Biological preparation: transgenic animals with cell-specific fluorescence. Transgenic animal models overexpressing fluorescent proteins in particular subsets of cells are available (The Jackson Laboratories, Bar Harbor, ME, USA). Fluorescent mice are chosen so that the cells of biological interest are labeled, for example, mice overexpressing YFP in motor neurons are bred with ALS transgenic mice, G93A hSOD1, to fluorescently label cells inducing ALS symptoms upon dysfunction (21). This method’s animal manipulations are approved by the Brandeis Animal Care and Use Committee and are carried out in the Brandeis University animal care facility in accordance with federal, local, and institutional guidelines. 2.2. Sensor Controlled Aerosol

1. ImagePrep (Bruker Daltonics, Billerica, MA, USA). 2. MALDI-TOF–TOF Ultraflex III mass spectrometer with 200 Hz solid-state laser (Bruker Daltonics, Billerica, MA, USA). 3. Scanning electron microscope Zeiss FESEM Supra55VP (Carl Zeiss SMT Inc., Peabody, MA, USA). 4. Confocal microscope Leica TCS SP2 AOBS (Acousto Optical Beam Splitter) Spectral Confocal Microscope equipped with a 405 nm UV laser (Leica Microsystems, Bannockburn, IL, USA). 5. Vectashield HardSet mounting media with 4′ ,6-diamidino2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA, USA). 6. Biological preparation: multimodal cell tracking using fluorescent dye uptake. Human U87 glioma cells are labeled by fluorescence and mass markers by in vitro uptake of DiI (λab 549 and λem 565 nm; m.w. 933.9 g mol−1 ) and human neural stem cells HB1.F3 are similarly labeled with DiO (λab 484 and λem 501 nm; m.w. 881.7 g mol−1 ). Vybrant DiI and DiO solutions are from Invitrogen

Tissue Preparation for the In Situ MALDI MS Imaging

419

(Carlsbad, CA, USA). Aliquots of 120,000 labeled U87 and HB1.F3 cells in 3 µl sterile PBS are surgically injected into the left brain hemisphere of anesthetized nude mice using a stereotactic device on days 1 and 10, respectively. Swiss nude male mice of 4−8 weeks of age (Charles River Laboratories, Wilmington, MA, USA) are anesthetized and killed by intraperitoneal 90 mg/kg ketamine and 10 mg/kg xylazine. Method 2 (Section 3.2) animal manipulations are committee approved and carried out in the animal facility at the Brigham and Women’s Hospital in accordance with federal, local, and institutional guidelines. Animals are sacrificed on day 14, and the brains readily dissected and flash frozen in liquid nitrogen. 7. Matrices: sinapinic acid and α-cyano-4-hydroxycinnamic acid (α-CHCA). 2.3. Microinjection with Matrix

1. Programmable Micropipette Puller, PMP-102 (MicroData Instrument Inc., Plainfield, NJ, USA). 2. Model PLI-100 Pico-Injector (Harvard Apparatus Inc., Holliston, MA, USA). 3. Microscope with mounted apparatus for microinjection. 4. Borosilicate glass capillaries (GC100-7.5) (Harvard Apparatus Part No. 30-0018, 1.0 mm OD×0.58 mm ID). 5. Micromanipulator with x, y, z control (Standard Manual Mechanical Micromanipulator, Harvard Apparatus Inc., Holliston MA, USA). 6. Biological preparation: same as in Method 1 (Section 2.1). 7. Rhodamine B – optional- aids in visualization.

3. Methods The methods presented herein are histology compatible protocols for use with either proteins (Methods 1 and 3) or smaller molecules, including peptides, lipids, and drugs and their metabolites (Methods 2 and 3). These protocols include matrix solution fixation (Section 3.1); a sensor controlled aerosol (Section 3.2), and matrix microinjection (Section 3.3), a novel method for labeling single cells. Matrix solution is optimized for proteins, prepares the entire surface of the tissue for imaging, results in tissue deformation and analyte redistribution on the order of ∼1 µm, and is among the most rapid of all matrix deposition methods. The sensor controlled aerosol we present is optimized for small

420

Agar et al.

(

H80

T4

05.0

H00

T5

L->

H75

T6

08.0

H00

T7

L->

H81

T8

0.6

H00

T9

L->

H00

T10

00.0

H00

PU1:02 PU1:02 PU1:01 PU2:05c

micropipette tip into the matrix solution and pressing the “fill” button. Allow the matrix solution to fill to half of the observable micropipette (the length of micropipette from tip to knurled nut) (see Notes 15 and 16). 4. Attach the combined hose and micropipette to a micromanipulator mounted on the microscope stage. Maintain an angle of ≥45◦ to prevent the matrix solution from going up the sides of the micropipette (see Note 17). Place a clean glass slide on the stage. The cover slip with tissue sample is placed onto this glass slide and taped to it to increase stability. 5. Inject cells under medium intensity light, and if fluorescently labeled cells are present, toggle back and forth between fluorescence and white light. Slowly move micropipette near target cell. If needed, increase light intensity in order to visualize the shadow of the micropipette. The shadow may be used as a guideline in order to line up the cell with the tip of the injector. Using the microscope x–y stage control, move the cell toward the tip of shadow. Now move the micropipette downward and toward the cell (see Note 5). When the tip of the micropipette touches the cell, a drop of matrix expands outward and encompasses the cell. This process may be repeated after the matrix solution solvent evaporates. Be sure to make a quick tap of the tip of micropipette onto the cell and remove promptly to avoid an overflow of matrix to surrounding cells (see Note 18).

426

Agar et al.

6. To visualize the injected cell with the MALDI mass spectrometer camera, it is necessary to delineate the injected area by scratching a contour around it with a dry micropipette. 7. Acquire mass (Fig. 24.4).

Relative intensity

A

spectra

from

single

cell

preparation

B

C

Fig. 24.4. MALDI mass spectrometry of a single neuron. (a) Positive ion mode mass spectrum of a single motor neuron from layer V of a mouse motor cortex. (b) Using a modified microinjection setup a single neuron was labeled with a mixture of the MALDI matrix, sinapinic acid, and Rhodamine B (for contrast in light microscopy and for visualizing with the MALDI microscope). Light microscopy is used to visualize deposition of the matrix and Rhodamine B. (c) The same microscopic field as above visualized using fluorescence microscopy. Microinjection of neurons is guided by neuron fluorescence. Mice are YFPH/WTSOD1 double transgenics with fluorescent motor neurons (described in methods).

4. Notes 1. The following fixative solution was optimized to maximize both quantity and quality of analyte MS signals from mouse brain tissue, while minimizing diffusion of proteins and tissue deformation. This solution was not optimized for lipids or peptides and indeed may solubilize and displace them. Users may substitute solvents or modify the ratio of solvents to optimize MALDI of molecules of interest and preserve tissue integrity. Ethanol, methanol, and acetone are among recognized fixatives, but independently yield limited peak detection, while water, TFA, and acetonitrile contribute to protein displacement, but increase

Tissue Preparation for the In Situ MALDI MS Imaging

427

extraction/detection. An optimized protocol provides a balance between analyte extraction and diffusion to favor matrix/analyte interface formation, while preserving the analyte’s spatial distribution. 2. Fresh matrix solution (less than 12 h old) and sonication (22) are critical for reproducible and homogeneous crystallization. 3. A readily available and reliable reference for standard histology methods, including cryosectioning, is the Online Information Center for Immunohistochemistry IHC World (www.ihcworld.com/). While thinner sections offer higher spectral quality, they are also more susceptible to analyte displacement. 4. Excessive thawing times decrease reproducibility (20). The thaw mounting procedure can be circumvented by placing a ∼3 µl droplet of matrix fixation solution on a cold target and then placing the tissue on this droplet. OCT suppresses the signal of tissue-derived analytes, so the frozen tissue is held to the sectioning support by a minimal amount of OCT at the base to hold the specimen in place during sectioning (4). Certain types of tissue may require full embedding for sectioning, so the excess OCT deposited on the glass is washed away with subsequent cold ethanol washes. 5. For smaller or larger tissue, use the Online Information Center at www.ihcworld.com to calculate the volume of tissue. Then use a ratio of fixing solvent to tissue of at least 20:1. 6. Drying temperature is a critical parameter. Drying temperature of 0 and 25◦ C (following 10 min at −22ºC) provides higher quality spectra but can result in analyte displacement. If users attempt to dry the matrix solution at a higher temperature (following 10 min fixing at −22ºC) care must be taken to keep the samples in dry air since condensation of atmospheric water on the cold surface changes the percentage of water in fixing solvent solution, thereby increasing analyte displacement. 7. The visual appearance of the matrix layer is a reliable indicator of its ionization/desorption efficiency potential. The most efficient matrix layers show a homogenous and shiny appearance whereas layers with a mat finish render lower quality spectral data. The laser spot size is estimated to be between 100 and 150 µm. 8. Sonication of the matrix solution, for any type of deposition, provides increased solution homogeneity and minimizes artifactual nucleation events. It is also recommended

428

Agar et al.

to centrifuge the solution after sonication to remove insoluble particles. 9. The use of high-resolution microscopy allows users to delineate the size of deposited vapor droplets and to resolve crystal arrangement, size, and homogeneity level. With such information, users can begin to correlate analytical characterization of matrix preparations with efficiency of MALDI. Characterization by SEM: The Supra55VP field emission scanning electron microscope (FESEM) allows surface examination down to nanometer scales either in high vacuum or in variable pressure (VP) mode. The instrument uses a low to moderate energy (0.1–30 keV) electron beam to image a sample with resolutions down to 1 nm at 15 keV in high vacuum or 2 nm at 30 keV in VP mode, which enables imaging of the matrix without coating. 10. Confocal microscopy provides molecular imaging with a higher spatial resolution than currently available for MSI, and protocols can be developed to minimize discrepancies in sample preparation for comparing both imaging methods. Confocal microscopy provides users with the ability to optically section a matrix-coated specimen and image a molecular target underneath a layer of matrix to assess the effects of sample preparations (see Fig. 24.2). Small molecular probes provide the opportunity for the direct detection of fluorescence and mass from the same molecule. The use of DAPI as a nuclear stain minimizes tissue manipulation as it is included in the mounting media and aids to guide the localization of a small population of fluorescent cells at high-resolution imaging of a large tissue specimen. 11. The tissue section was imaged for masses between 500 and 3,100 m/z, and the distribution of mass 781.7 ± 0.2 m/z, which corresponds to the calculated monoisotopic mass of a +1 charge of DiO in solution. The spatial distribution of the signal corresponds to the expected gross anatomical distribution of the stereotactically injected stem cells and to confocal microscopy results. 12. The isolated fluorescence signal indicated with an arrow in Fig. 24.3a is questionable as it is distant from the injection site, and tissue typically shows some level of autofluorescence. Even though stem cells can migrate away from the injection site, the presentation of the fluorescent signal is validated by in situ fragmentation of the 781.7 m/z signal with comparison to standard DiO’s fragmentation pattern. In this case, the use of MSI offers greater certainty to the visualization of migrating stem cells than confocal microscopy, illustrating the potential of mass spectrometry as an imaging tool.

Tissue Preparation for the In Situ MALDI MS Imaging

429

13. Spectra from a defined region of interest in contralateral regions of the brains, including the fluorescent cells injection site, are saved and used to build classification models in ClinProTools (Bruker Daltonics, Billerica, MA, USA) with genetic algorithm (GA; 96% recognition capability), single neural network (SNN; 94%), and quick classifier (QC; 90%). Four spectra from a third region of the brain showing weak 781.7 m/z intensity are classified and recognized as human neural stem cells for 4/4 spectra with both the GA and QC, and 2/4 spectra with the SNN derived model. 14. Try to produce a micropipette with a long taper, which will prevent the tip from breaking too easily. Considerable variation in pipette shape occurs even under identical instrument conditions. Variation depends upon the age of the heating filament, the pressure of the “pressure 1” parameter, and other minor factors (see manual of PMP102). To maintain pipette shape, alter the heat level of T7 or change the time of T8. A higher heat of T7 will produce a smaller tip, a longer time of T8 will produce a larger tip. 15. If matrix does not move into the micropipette, the tip may be too small. To increase tip size, decrease heat level of T7 or increase operating time on T8. 16. The presence of small volumes of matrix solution in the micropipette helps prevent the matrix from flowing back into the tip. 17. If the micropipette scrapes other cells when being pulled back, either the micropipette tip is too long or the micropipette has been placed at too shallow of an angle. Adjust for a larger angle between micropipette and stage or (while retracting micropipette) move slide toward direction in which micropipette is retracting toward. 18. If droplets of matrix form on the tip, either the balance pressure is too high, the internal diameter of the tip is too large, or air is leaking in from the knurled nut. If decreasing the balance pressure does not remediate the problem, replacing the silicon rubber gasket might (Digitimer, Ltd. Hertfordshire, England Model PLI-SRG-1.0. Silicone Rubber Gasket Replacement for use with PLI-PH1 and PLI-PH1A). Alternatively, use a wide and thin piece of parafilm (Parafilm (M) Laboratory film, Fisher Inc.) to wrap around the area of contact between the micropipette and the knurled nut in order to make it airtight. 19. If after a few injections, matrix does not come out of the micropipette, it is likely that some crystallization has

430

Agar et al.

occurred on the tip of the micropipette. You can either take a piece of wet paper to scrape the sides of the tip without breaking it or slightly increase the balance pressure (0.01–0.07 psi or more) to force the matrix out.

Acknowledgments This work was made possible by award W81×WH-04-0158 from the Department of Defense to JA; grant 1392 from the Amyotrophic Lateral Sclerosis Society of America to JA; Brain Science Foundation to NYRA; American Brain Tumor Association to NYRA, and the Daniel E. Ponton Fund for the Neurosciences to NYRA. We acknowledge the Brandeis University Mass Spectrometry Resource, and the Brandeis University Animal Care Facility for care of instruments and animals, respectively. This work was performed in part at the Center for Nanoscale Systems (CNS), a member of the National Nanotechnology Infrastructure Network (NNIN), which is supported by the National Science Foundation under NSF award no. ECS-0335765. CNS is part of the Faculty of Arts and Sciences at Harvard University. We also wish to thank Ed Dougherty for maintenance and support of optical microscopy at Brandeis University, Drs. Sacha Nelson and Ken Sugino for transgenic mice, and we also wish to thank Dr. Lata G. Menon for animal surgery and manipulations. References 1. Gusev, A. I., Vasseur, O. J., Proctor, A., Sharkey, A. G., Hercules, D.M. (1995) Imaging of thin-layer chromatograms using matrix-assisted laser desorption/ionization mass spectrometry. Anal Chem, 67, 4565–4570. 2. Aerni, H. R., Cornett, D. S., Caprioli, R. M. (2006) Automated acoustic matrix deposition for MALDI sample preparation. Anal Chem, 78, 827–834. 3. Sloane, A. J., Duff, J. L., Wilson, N. L., Gandhi, P. S., Hill, C. J., Hopwood, F. G., Smith, P. E., Thomas, M. L., Cole, R. A., Packer, N. H., et al. (2002) High throughput peptide mass fingerprinting and protein macroarray analysis using chemical printing strategies. Mol Cell Proteomics, 1, 490–499. 4. Schwartz, S. A., Reyzer, M. L., Caprioli, R. M. (2003) Direct tissue analysis using matrix-assisted laser desorption/ionization

5.

6.

7.

8.

mass spectrometry: practical aspects of sample preparation. J Mass Spectrom, 38, 699–708. Jurchen, J. C., Rubakhin, S. S., Sweedler, J. V. (2005) MALDI-MS imaging of features smaller than the size of the laser beam. J Am Soc Mass Spectrom, 16, 1654–1659. Puolitaival, S. M., Burnum, K. E., Cornett, D. S., Caprioli, R. M. (2008) Solvent-free matrix dry-coating for MALDI imaging of phospholipids. J Am Soc Mass Spectrom, 19, 882–886. Lemaire, R., Wisztorski, M., Desmons, A., Tabet, J. C., Day, R., Salzet, M., Fournier, I. (2006) MALDI-MS direct tissue analysis of proteins: improving signal sensitivity using organic treatments. Anal Chem, 78, 7145–7153. Agar, N. Y., Yang, H. W., Carroll, R. S., Black, P. M., Agar, J. N. (2007) Matrix

Tissue Preparation for the In Situ MALDI MS Imaging

9. 10. 11. 12. 13.

14.

15.

16.

solution fixation: histology-compatible tissue preparation for MALDI mass spectrometry imaging. Anal Chem, 79, 7416–7423. Clarke, J. L. (1851) Researches into the structure of the spinal cord. Phil Trans Roy Soc, 141, 601–622. Carnoy, J. B. (1887) La cytodierese de l′ oeuf. Cellule, 3, 6. Blum, F. (1893) Der Formaldehyde als Hartungsmittel. Vorlaufige Mittheilung. Z.w.M., 10, 314–315. Prausnitz, J. M. (2003) Molecular thermodynamics for some applications in biotechnology. Pure Appl Chem, 75, 859–873. Xu, B. J., Li, J., Beauchamp, R. D., Shyr, Y., Li, M., Washington, M. K., Yeatman, T. J., Whitehead, R. H., Coffey, R. J., Caprioli, R. M. (2009) Identification of early intestinal neoplasia protein biomarkers using laser capture microdissection and MALDI MS. Mol Cell Proteomics, 8, 936–945. Neupert, S., Johard, H. A., Nassel, D. R., Predel, R. (2007) Single-cell peptidomics of drosophila melanogaster neurons identified by Gal4-driven fluorescence. Anal Chem, 79, 3690–3694. Monroe, E. B., Jurchen, J. C., Koszczuk, B. A., Losh, J. L., Rubakhin, S. S. Sweedler, J. V. (2006) Massively parallel sample preparation for the MALDI MS analyses of tissues. Anal Chem, 78, 6826–6832. Holle, A., Haase, A., Kayser, M., Hohndorf, J. (2006) Optimizing UV laser focus profiles

17.

18.

19.

20.

21.

22.

431

for improved MALDI performance. J Mass Spectrom, 41, 705–716. Spengler, B., Hubert, M. (2002) Scanning microprobe matrix-assisted laser desorption ionization (SMALDI) mass spectrometry: instrumentation for sub-micrometer resolved LDI and MALDI surface analysis. J Am Soc Mass Spectrom, 13, 735–748. Luxembourg, S. L., Mize, T. H., McDonnell, L. A., Heeren, R. M. (2004) Highspatial resolution mass spectrometric imaging of peptide and protein distributions on a surface. Anal Chem, 76, 5339–5344. Luxembourg, S. L., McDonnell, L. A., Mize, T. H., Heeren, R. M. (2005) Infrared mass spectrometric imaging below the diffraction limit. J Proteome Res, 4, 671–673. Goodwin, R. J., Dungworth, J. C., Cobb, S. R., Pitt, A. R. (2008) Time-dependent evolution of tissue markers by MALDI-MS imaging. Proteomics, 8, 3801–3808. Schaefer, A. M., Sanes, J. R., Lichtman, J. W. (2005) A compensatory subpopulation of motor neurons in a mouse model of amyotrophic lateral sclerosis. J Comp Neurol, 490, 209–219. Lemaire, R., Tabet, J. C., Ducoroy, P., Hendra, J. B., Salzet, M., Fournier, I. (2006) Solid ionic matrixes for direct tissue analysis and MALDI imaging. Anal Chem, 78, 809–819.

Chapter 25 Imaging of Similar Mass Neuropeptides in Neuronal Tissue by Enhanced Resolution MALDI MS with an Ion Trap – OrbitrapTM Hybrid Instrument Peter D.E.M. Verhaert, Martijn W.H. Pinkse, Kerstin Strupat, and Maria C. Prieto Conaway Abstract Several mass spectrometry imaging (MSI) procedures are used to localize physiologically active peptides in neuronal tissue from American cockroach (Periplaneta americana) neurosecretory organs. We report how to use this model system to assess, for the first time, the performance of the MALDI LTQ OrbitrapTM XL mass spectrometer to perform MSI of secretory neuropeptides. The method involves the following steps: (1) rapid dissecting of neurosecretory tissue (i.e., insect neurohemal organ) in isotonic sucrose solution; (2) mounting the tissue on a glass slide; (3) controlled spraying of the air-dried tissue with concentrated MALDI matrix solution; (4) loading specimen into the MALDI source of a MSn system equipped with an OrbitrapTM analyzer; (5) setting-up MSI methods by determining tissue areas of interest, spatial resolution, molecular mass range, and molecular mass resolution; (6) acquiring mass spectra; (7) analyzing data using ImageQuestTM MSI software to generate (single or composite) images of the distribution of peptide(s) of interest; (8) confirming the identity of selected peptides by MS2 and/or MSn sequencing directly from imaged tissue sample. The results illustrate that high mass accuracy and high mass resolving power of the Orbitrap analyzer are achievable in analyses directly from tissue, such as in MSI experiments. Moreover the mass spectrometric instrumentation evaluated allows for both peptide localization and peptide identification/sequencing directly from tissue. Key words: Cockroach, mass spectrometry imaging, neuropeptides, Orbitrap detector.

1. Introduction There is great interest in mass spectrometry imaging (MSI) of secretory peptides, as they are essential molecules for different physiological processes, comprising a major portion of the cellular S.S. Rubakhin, J.V. Sweedler (eds.), Mass Spectrometry Imaging, Methods in Molecular Biology 656, DOI 10.1007/978-1-60761-746-4_25, © Springer Science+Business Media, LLC 2010

433

434

Verhaert et al.

(bio)chemical “language.” Whereas the overall fate/tendency of signaling compounds is to eventually end up in the intercellular space or in the general circulation (which makes their detection and visualization at the site of release quite a challenge), it is of great interest to be able to trace the source of these biochemicals. An obvious example is the discovery of potential biomarkers within their site of synthesis, e.g., diseased/malignant tissue, such as a specific tumor. For obvious methodological and ethical reasons it is unreasonable to obtain fresh unfixed human tumor tissue for the sole purpose of MSI method development. Additionally, tumor biopsies typically come in a wide variety of amorphous shapes, which make proper (immuno)histochemical controls for each sample a must to validate developed MSI protocols. Frequently utilized in such projects are samples from rodent (mouse or rat) brains or pituitaries. However, the sizes and shapes of these tissues require cryotome sectioning, which produces 5–15 µm thick tissue slices amenable for MALDI matrix application and MSI interrogation. Sectioning typically results in significant increase in the duration of sample preparation and the need for freeze and thaw of specimens. To circumvent these drawbacks, we have been focusing on the cockroach retrocerebral neurosecretory system, which has proven to be a truly elegant biological model for evaluating novel investigative approaches in secretory peptide MSI. It is small enough to allow “cryosection free” and relatively quick high spatial resolution tissue analysis, while being large enough to allow facile dissection from the animal. In evolutionarily ancient insects such as cockroaches, the neuroendocrine glands consist of two pairs of distinctly shaped organs. By their mere shape (Fig. 25.1), one easily distinguishes the elongated corpora cardiaca (cc) with a cranial, somewhat thicker glandular part (ccg) and a more caudal storage part (ccs) and the oval/round-shaped corpora allata (ca). The insect cc and ca comprise the animal’s major neurosecretory tissue, equivalent with the pituitary or hypophysis gland of higher vertebrates. They are known to contain a great variety of (neuro)secretory peptides, many of which have had their primary sequences elucidated over the past 20 years, with the American cockroach, Periplaneta americana, being one of the most extensively studied species (Table 25.1). Typical for arthropod (such as insects, spiders, crustaceans) neurosecretory tissue is that the secretory part of the glands is the outside surface (see Fig. 25.1b; 1), which, therefore, is readily accessible for MALDI matrix application. Furthermore, the tiny size of the tissue allows for whole mount tissue analysis (i.e., without the need of a cryostat) and for full gland analysis within a relatively short time frame. An additional advantage is that live

Imaging of Similar Mass Neuropeptides in Neuronal Tissue

435

B

A ca ccs

ccg 1000 μ m

D ccg

C

ccs ccg

ca

ccs

Fig. 25.1. American cockroach retrocerebral complex. (a) Tissue complex as dissected out of the animal head; preparation in 250 mM sucrose solution showing the pairwise structure of the glands with cranial glandular lobes of the corpora cardiaca (ccg) linked with the corpora allata (ca) via the storage lobe of the cc (ccs). (b) Scanning electron micrograph of rapidly fixed whole mount cc showing (neuro)secretion (white “stuff”) taking place at the outer surface (ccs part; picture reproduced from (1) with permission). (c, d) Different histological sections; (c) Light microscopic view (hematoxylin and eosin stain) of cc and ca showing glandular (ccg) and storage lobe (ccs); (d) Classic immunohistochemistry stain (monoclonal neuropeptide antibody in immunoperoxidase technique, see (17)), showing neuropeptide containing nerve fibers in the center of the tissue (arrows) which are extensively branching in the gland periphery, particularly in the storage lobe (ccs).

cockroaches are omnipresent throughout the globe. They can be readily obtained virtually everywhere (if not in “the wild,” then through purchase at local “pet shops” or “biologicals for education” suppliers). All these features have made the cockroach neurosecretory system our favorite model for direct tissue (neuro)secretory peptide investigations, for many years (2). It is obvious that the final quality of MSI data largely, but not solely, depends on the tissue preparation and processing steps. Also, the capabilities and performance of the mass spectrometer used are of importance. Over the past 5 years we have been optimizing a direct tissue (neuro)peptide sample preparation method, in combination with the use of various mass spectrometers, including MALDI time-of-flight (TOF), MALDI-TOF– TOF, MALDI ion traps, MALDI Q TOF, and others (2–5). Here we report the MSI protocol based on a combination of our optimized sample preparation and an ion trap – Orbitrap hybrid mass analyzer (6) sample interrogation, followed by data processing and the creation of molecular images of neuropeptide distributions.

436

Verhaert et al.

Table 25.1 Known P. americana (Pea) cc/ca neuropeptide sequences and accurate monoisotopic (m.i.) masses calculated for their ions as expected (i.e., observed in earlier direct tissue MS analyses (9, 10)). Calculations of theoretical peptide masses are done using protein calculator m.i. Mass [M+H]

(→ Derived masses)

Sequence (a )

[Trivial name]

SPPFAPRLamide

[Pea-PK-II]

883.514848

pQVNFSPNWamide

[Pea-CAH-I]

973.452642

(→ 995.434586 [M+Na], 1011.408524 [M+K])

pQLTFTPNWamide

[Pea-CAH-II]

988.488693

(→ 1010.470637 [M+Na], 1026.444575 [M+K])

LVPFRPRLamide

[Pea-PK-III]

996.646531

HTAGFIPRLamide

[Pea-PK-I]

1010.589410

pQDVDHVFLRFamide

[Pea-LMS]

1257.637482

FDDY(SO3)GHMRFamide

[